Research on Precision Marketing Model of Tourism Industry

Research ArticleResearch on Precision Marketing Model of Tourism IndustryBased on Userrsquos Mobile Behavior Trajectory

Jialin Zhang12 Tong Wu1 and Zhipeng Fan 12

1Harbin University of Commerce Harbin 150028 China2Heilongjiang Provincial Key Laboratory of Electronic Commerce and Information ProcessingComputer and Information Engineering College China

Correspondence should be addressed to Zhipeng Fan hsdfzp126com

Received 28 September 2018 Revised 4 December 2018 Accepted 13 December 2018 Published 3 February 2019

Guest Editor Jaegeol Yim

Copyright copy 2019 Jialin Zhang et al is is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

With the deep cross-border integration of tourism and big data the personalized demand of tourist groups is increasingly strongPrecision marketing has become a new marketing mode that the tourism industry needs to pay close attention to and exploreBased on the advantages of big data platform and location-based service starting from the precise marketing demand of tourismwe design data flow mining technology framework for userrsquos mobile behavior trajectory based on location services in mobilee-commerce environment to get user track data that incorporates location information consumption information and socialinformation Data mining clustering technology is used to analyze the characteristics of usersrsquo mobile behavior trajectories andthe precise recommendation system of tourism is constructed to provide support for tourism decision making It can target thetourist group for precise marketing and make tourists travel smarter

1 Introduction

11 Research Background Location-based service (LBS) is akind of value-added service provided by the combination ofmobile communication network and satellite positioningsystem It obtains the location information of mobile terminalsuch as latitude and longitude coordinate data through a set ofpositioning technology and provides it to the communicationsystem mobile users and related users to realize variouslocation-related services in military and transportation As anew mobile computing service in recent years 80 of theworldrsquos information has time and location tags and locationservices have developed to the big data stage [1] Developinglocation services requires two capabilities the ability toprovide location and the ability to understand location

e precise marketing information pushed by LBS lo-cation service can effectively tap the potential consumerdemand and make a scientific and reasonable networkmarketing strategy based on this which can further improvethe ability of e-commerce enterprises to tap the target cus-tomers and potential customers According to the 2017 ChinaMobile e-commerce Industry Research Report the transaction

scale of Chinarsquos e-commerce market reached 202 trillion yuanin 2016 an increase of 236 compared with the same period ofthe same period of the year Chinarsquos e-commerce market isdeveloping steadily Among them the online shopping has agood momentum of development up from 233 in 2015Huge market potential tempts all walks of life In 2016 onlineshopping and B2B e-commerce of small and medium-sizedenterprises and enterprises above scale still dominate theChinese e-commerce market while online tourism and locallife service O2O emerge as bamboo shoots accounting for3 and 16 of the market respectively From 2015 to2016 the proportion of online tourism market in Chinarsquostourism market has greatly increased the process ofproduct informatization has accelerated the penetrationrate has further improved the mobile online tourismmarket has developed rapidly and consumer un-derstanding and demand and experience of tourism arechanging imperceptibly and pursuing higher quality oftourism With the further development of ldquoInternet+rdquoinformation technology the tourism industry has hugeroom for development and online travel penetration willalso gradually increase

HindawiMobile Information SystemsVolume 2019 Article ID 6560848 14 pageshttpsdoiorg10115520196560848

According to the Statistical Bulletin on National Eco-nomic and Social Development of 2016 issued by the StateStatistical Bureau in the whole year of 2016 the number ofdomestic touristsrsquo trips reached 44 billion an increase of112 over the previous year and the income of domestictourism increased by 152 to 39390 billion yuan enumber of inbound tourists reached 13844 million anincrease of 35 and international tourism revenue in-creased by 56 to $120 billion e number of domesticresidents in China has reached 135 million 130 thousand anincrease of 57 [2] With the continuous promotion of thestrategic pace of building a well-off society in an all-roundway tourism has become an important part of the peoplersquosdaily life in China marking that Chinarsquos tourism industryhas entered the era of mass tourism

Based on this background in the mobile e-commerceenvironment based on LBS location service research andanalysis of userrsquos mobile behavior trajectory can extractvaluable userrsquos mobile behavior features from a largenumber of mixed dynamic data and integrate the mobilebehavior and consumer behavior of tourism users Basedon LBS location services will integrate the mobile be-havior and consumer behavior of tourism users thenexcavate the marketing value of consumers and timelyachieve the marketing objectives of enterprises on theappropriate media so that mobile e-commerce marketingbecomes more accurate and effective rough the re-search of this subject the interests of enterprises con-sumers and media can be maximized at the same timeproviding personalized products and services for mobilee-commerce improving consumer loyalty and corecompetitiveness of mobile e-commerce and bringinghigher profits for e-commerce enterprises [3]

12 Presentation of Problems e data of mobile terminalusersrsquo historical consumption behavior and locationmovement process are recorded and stored according tothe time series forming the userrsquos mobile behavior tra-jectory data which can be collected by multiple deviceterminals e userrsquos mobile behavior trajectory datacontain a lot of useful information Mobile behavior tra-jectory can express the behavior activities of mobile users inthe real world ese activities imply userrsquos interestshobbies experiences and behavior patterns [4] For ex-ample a userrsquos activities in a week may start from home towork every day and a user may go to shopping malls parksand other places on weekends erefore how to effectivelyutilize the userrsquos mobile behavior trajectory and extractuseful information from the userrsquos mobile behavior tra-jectory data is very important for the realization of per-sonalized recommendation service

From the view of consulting a large number of docu-ments there are more papers on location services than onmobile marketing However most of the previous articles onlocation services focused on the application of natural sciencesuch as surveying and mapping technology network devel-opment geographic information and so on In recent yearsthe number of cross-research articles onmanagement science

medicine and agriculture combined with location-basedservices has begun to increase most of which are the com-bination of location-based services and related industries tostudy the application or technology development of specificindustries For example the combination of location servicesand logistics technology can track the journey of packagesCombining location services with electronic maps can pro-vide catering entertainment discounts and other in-formation within a certain range according to the location ofusers Combining location service with utility technology canquickly find information such as tap water gas explosion andso on e main keywords of literature research include lo-cation service technology location service system locationservice terminal location service strategy mobile locationservice and so on Based on location the services industry iscurrently considered one of the most dynamic industriesWith the rapid development of mobile Internet and Internetof ings technology more debris time has been transferredto mobile phones tablets and smart products [5]

e characteristics of mobile marketing such as pre-cision interaction novelty and effective delivery are moreand more concerned and recognized by various industriesis paper focuses on the main characteristics of the end-users of the tourism industry such as frequent locationmovement strong sense of sharing and rich demand forservices First of all the users of online travel must bemobile users who usually do not stay in a location for a longtime and a high probability of frequent location changeswill produce a large number of location data Moreover ingeneral traveling users arrive at an unknown location ormove in a series of unfamiliar geographical environmentswhich makes travel usersrsquo demand for location-basedservices take precedence over personal privacy pro-tection and enable them to obtain real-time user locationinformation ese location data provide favorable con-ditions for our research Secondly the behavior of touristusers is quite different from that of ordinary people Inbeautiful scenery and not very familiar environment userswill spontaneously produce self-awareness Most peopleshare location photos and moods through social platformsandmicromessaging and travel companies can access thesesocial data to accurately portray users and provide accurateservices for them ird travel users need high-qualityservices to obtain high-quality tourism experience Sce-nic spots accommodation restaurants transportation andfinance including tour guides and their fellow travelers arealso important factors in achieving a high-quality tourismexperience ese rich demands for services will generateenormous commercial value [6] erefore this articleadopts top-down overall analysis to design ideas frombottom to top By analyzing the userrsquos characteristicsthrough the trajectory of userrsquos mobile behavior this paperconstructs a travel recommendation system in the mobilepoint-to-point environment and a precise marketing modelin the tourism industry based on the trajectory of userrsquosmobile behavior so as to provide appropriate services forthe appropriate users at the right time and place in order toprovide reference for relevant tourism enterprises toachieve precise marketing

2 Mobile Information Systems

2 Userrsquos Mobile Behavior Trajectory

21 Userrsquos Mobile Behavior Trajectory Definition Userrsquosmobile behavior trajectory is based on the path that users findfrequently in the location mobile path generated by daily lifee location information generated by userrsquos daily behavior isacquired by GPS equipment sampling at a certain time in-terval and the spatial position of moving object is representedby Euclidean space coordinates discrete display in electronicmap rough moving sequence pattern mining we can findthe correlation among these discrete location informationpoints and obtain the userrsquos moving behavior trajectory iswill provide effective support for precision marketing inmobile e-commerce [7] In this paper we make the followingdefinitions for userrsquos mobile behavior trajectory

Definition 1 Location information point the position in-formation points generated by the userrsquos movement can beobtained by receiving devices such as GPS of mobile ter-minals Each position information point indicates a positionthat the user has arrived at Suppose an independent locationinformation point is represented as two tuple P (Z T)among them Z is the position coordinate and its structurecontains longitude Z x and latitude Z y T is the time in-formation of arrival position Z

Definition 2 Mobile behavior trajectory mobile behaviortrajectory can be obtained by GPS log A mobile behaviortrajectory consists of a sequence of position infor-mation points arranged in order of time attribute TSuppose L is userrsquos mobile behavior trajectory thenL P1⟶ P2⟶ middot middot middot⟶ Pn where Pi(0lt ile n) denotesany sampled position information point Mobile behaviortrajectory L satisfies any 0lt ile n Pi middot TltPi+1 middot Ta n rep-resents the number of location information points andrepresents it as the length n of mobile behavior trajectory

Definition 3 Mobile behavior subtrajectory represents theinclusion or inclusion relationship between twomoving behavior trajectories Suppose that L1 and L2have two trajectories of moving behavior where L1

a1⟶ a2⟶ middot middot middot⟶ ai L2 b1⟶ b2⟶ middot middot middot⟶ bnIf there exists a positive integer m1 m2 mi satisfying1lem1 ltm2 lt middot middot middot ltmi le n making a1 bm1 a2 bm2

ai bmi then L1 is said to be the moving behavior sub-trajectory of L2 or L2 is said to be a moving behaviorsupertrajectory of L1 It can be written as L1 sube L2 or L2 supe L1e location information points are adjacent to the mobilebehavior subtrajectory and are allowed to be nonadjacent inthe original mobile behavior trajectory

Definition 4 Support degree the collection of all locationinformation points of moving behavior trajectoriesconstitute a database of mobile behavior trajectoriesDB L1 L2 L3 Ln1113864 1113865 where Li(0lt ile n) is mobile be-havior trajectory and |DB| is the number of mobile behaviortrajectories in the database e number of mobile behaviortrajectory t contained in DB is t of the support in DB

support Li( 1113857 L | L isin D Li sube L1113864 11138651113868111386811138681113868

1113868111386811138681113868 (1)

Definition 5 Frequent Trajectories when the supportdegree of the mobile behavior trajectory is greater than orequal to the minimum support threshold the mobilebehavior trajectory is called the frequent trajectoryFT l | support(l)gemin lsube L L isin D1113864 1113865 L represents themobile behavior trajectory sequence and D represents themobile behavior trajectory sequence set

e userrsquos mobile trajectory records the userrsquos activitystatus in the real world which can reflect the userrsquos behaviorpreferences and potential intentions to some extent Forexample if a user moves a lot every day he may be anoutdoor sports enthusiast rough more fine-grainedanalysis we can identify usersrsquo occupations taste habitsand so on from their frequent locations and restaurantserefore mining hot spots and planning roads throughmultiuser mobile trajectory data sharing is an importantresearch content of this paper

22 Classification of Userrsquos Mobile Behavior TrajectoryUserrsquos mobile behavior trajectory data refer to the sequenceof changes in geographic location information caused byuserrsquos own motion behavior in a certain time and spaceenvironment ese geographic location information pointswhich change with time series can form a userrsquos mobilebehavior trajectory data according to the order of occurrencetime [8] According to the different sampling methods wecan classify these userrsquos mobile behavior trajectory data intothree categories

221 Location Sampling-Based Userrsquos Mobile BehaviorTrajectory A trajectory formed by a change in positionduring the movement of a user can be sampled sequentiallyaccording to the change in position It focuses on the in-formation of location change when the user moves e dataobtained by this method have abundant semantic in-formation and very detailed location change informationWe can record the trajectory data of userrsquos mobile behaviorbased on position sampling by recording discrete variablese trajectory of userrsquos mobile behavior can be representedby the sequence of sampling points with the change ofmoving objectrsquos position and it can be formally expressed as

L x1 y1 t1 ( 1113857 xi yi ti ( 1113857 xn yn tn ( 11138571113864 1113865

(2)

e location (xi yi) 1le ile n denotes the geographicallocation of the mobile user at the time of ti and the location(xi yi) of the mobile user at the time of ti and the location(xi+1 yi+1) of the time of ti+1 are not the same

Trajectory can be divided into three segments accordingto the information of stopping point boarding point andalighting position and the trajectory can be preservedaccording to different semantics and application segmentsFor example in the prediction of travel time it is necessary

Mobile Information Systems 3

to delete the stopping point which may be the vehicleparking or waiting for passengers in order to measure thetrajectory travel time more accurately For some tasks thatanalyze the similarity between two users it is often necessaryto use the residence trajectory to reflect the userrsquos region ofinterest

222 Time Sampling-Based Userrsquos Mobile BehaviorTrajectory e change of mobile userrsquos behavior is sampledby definite time interval to form the trajectory data of userrsquosmobile behavior which is called the trajectory of userrsquosmobile behavior sampled according to time is kind ofsampling focused on the change of location informationpoints caused by the change of mobile userrsquos behavior at thesame time interval which has the characteristics of large datavolume and wide range e time-sampled trajectory data ofuserrsquos mobile behavior is formalized as follows


ti t1 +(iminus 1)Δt(3)

where L is a trajectory data of mobile behavior Δt is equalinterval time (xi yi) and 1le ile n denotes the location ofthe mobile user at any time of ti If the time interval betweenthe two sampling points is larger than the threshold valuethe trajectory can be divided into two segments through thetwo sampling points

223 Userrsquos Mobile Behavior Trajectory Triggered by Eventse trajectory of mobile userrsquos mobile behavior which isrecorded by the system after the sensor event is triggered isobtained by the event triggering [9] is sampling methodfocuses on the event set that triggers the sensor to work Ithas the characteristics of short update period and repre-sentative sampling objects Although the behavior of mobileusers changes with time the system does not record thetrajectory according to time or position but only records thetrajectory information of mobile users when they producesome specific behavior and trigger sensor events We canalso use discrete variables to record the behavior trajectory ofmobile users and formalize it as follows


(4)

e location (xi yi) 1le ile n denotes the location of themobile user at the time of ti and the location of the mobileuser at the time of ti (xi yi) and ti+1 can be the same(xi+1 yi+1)

When the trajectory direction changes beyond thethreshold value we can mark the key points according to thedirection changes and divide the trajectory into two segments

23 Userrsquos Mobile Behavior Pattern Decision According tothe trajectory data of userrsquos movement behavior the speed ofcompleting the trajectory is calculated by time and then theuserrsquos behavior pattern is determined Many problems still

need to be considered such as road congestion construc-tion and even traffic accidents which will affect the speed ofuser behavior Vehicles travel much faster than peoplersquoswalking speed on normal roads but in congested or ab-normal roads the speed difference between vehicles andpeoplersquos walking speed is not obvious erefore theidentification accuracy of trajectory velocity can only be lessthan 50 through time calculation [10] In addition the usermay change several different behavior patterns in the sametrip which makes the same userrsquos moving behavior trackcontain a variety of different speeds In the overall calcu-lation if the average speed is obtained it is obviously notcorrect to determine the userrsquos behavior patterns ereforeit is necessary to divide the userrsquos moving behavior trajectoryinto several trajectory segments reasonably By comparingdifferent trajectory segments we can analyze whether theuser has changed the behavior pattern and further improvethe recognition accuracy

How to realize the reasonable division of user movementbehavior segments is the problem we want to study Asshown in Figure 1 the walking user and the driving usertravel the same way but the trajectory data of the userrsquosmovement behavior are obviously different We can analyzethe following three aspects

(1) Because the trajectory data of userrsquos moving behaviorproduced by walking often produce direction changeor reciprocating motion we can divide the trajectorysegments according to the change of the trajectory datadirection of userrsquos moving behavior In mobile scenespeople get off a bus walk to another station to con-tinue to take the bus process and must pass through asection of walking although the walking section isshort but still can show obvious direction changes

(2) e trajectory data of mobile behavior produced bydriving users do not change significantly in directionis kind of characteristic is not affected by trafficconditionsWe can train a classificationmodel by thesupervised learning method For example drivers donot change their direction as freely and frequently aspedestrians do resulting in a straight line in thetrajectory of the userrsquos movement behavior and thedirection of change is not obvious [11]

(3) We can also judge user behavior patterns by theshape of user behavior trajectory data especially thetrajectory of user behavior generated by differentuser behavior patterns in a journey which will haveobvious morphological changes of trajectory

3 Analysis of Userrsquos Mobile BehaviorTrajectory Data

is paper studies the trajectory of userrsquos mobile behaviorgenerated by online travel users during their mobile processIt contains a lot of information to express the personalizedbehavior of mobile users We can use data mining methodssuch as classification clustering frequent itemsets cyclediscovery and anomaly detection to mine and analyze thetrajectory of tourism usersrsquo mobile behavior


31 Dividing Trajectory Segments Each user movementbehavior trajectory can be regarded as an image dataStructural Similarity Index (SSIM) can effectively measurethe similarity of two trajectories and clustering based on thesimilarity index is more accurate than traditional clusteringbased on Euclidean distance index [12] e accuracy ofstructure similarity matching is closely related to the rea-sonableness and validity of the segmentation of user motiontrajectory erefore this section mainly studies how todetect the large-angle mutation points in the userrsquos movingbehavior trajectory and how to partition and store the userrsquosmoving behavior trajectory records at the mutation pointsso as to obtain some trajectory fragments which tend to bestable before clustering

Each user movement behavior trajectory cannot be astraight line As the precision of position coordinate re-cording is higher and higher the direction of each track willchange more and more especially some subtle directionchanges and the angle of rotation can reflect the degree ofchange of the track direction e division of track segmentsis determined according to the size of the track angleHowever if every corner is stored it is not conducive toreduce the storage of the corner and it is not conducive toextract it to divide the trajectory segments erefore bystoring the large turning point we can discover and identifythe changes of user behavior or abnormal conditions whichis also conducive to retaining the relatively stable localstructure features of user trajectory segments

We define the turning angle of userrsquos moving behaviortrajectory as the turning angle caused by the change of di-rection of adjacent trajectory segments which can reflect themovement trend of trajectory and the change of userrsquos be-havior [13] As shown in Figure 2 the angle between thedirection changes of the userrsquos moving behavior trajectory canbe expressed as α and the angle of rotation can be divided intoouter angle and inner angle expressed as θ1 and θ2 re-spectively We set the outer rotation angle θ1 as a positivevalue and the inner rotation angle θ2 as a negative value tofacilitate the similarity calculation of the trajectory segments

As can be seen from Figure 2 the formula for calculatingthe angle alpha of the direction change is shown in Formula(5) where a b and c represent the adjacent and oppositesides of the angle α respectively

a arccosa2 + b2 minus c2

2ab (5)

According to the above formula the formula for cal-culating the angle theta can be obtained (6)

θ 180minus α if(a times bge 0)

αminus 180 if(a times blt 0)1113896 (6)

is is the first step to partition the trajectory segmentsof userrsquos mobile behavior Using formulas (5) and (6) thetrajectory segment partitioning algorithm can be imple-mented (Algorithm 1)

Some trajectory fragments obtained by calculatingrotation angle setting threshold and partitioning trajec-tory fragments can be expressed as a set of several featureattribute vectors ese feature attributes can compre-hensively express the local features of a trajectory fragmentand the global features of userrsquos moving behavior trajec-tory In this section the trajectory fragment is not simplythe expression of coordinate information of the positioninformation points but extracts the speed shape positionrotation angle acceleration and other characteristic vec-tors from it Using these eigenvectors we can enhance theaccuracy of analyzing the trajectory of user movement Weformally represent the trajectory fragment structure asfollows TS (D S A L) In addition to the above fourfeatures we should also calculate the distance time andother features using vector W WD WS WA WL1113864 1113865 torepresent the weight of the four feature vectors

Since the weights of feature vectors correspond to theeigenvectors of the trajectory segments their values shouldbe greater than or equal to zero and the sum of their weightsshould be 1 we can generally assume that the weights of allfeature vectors are equal probability and we can take theaverage value of 025 as the weights Similarly we can adjustthe weights of each feature vector according to the sensitivityof the feature vectors of the trajectory fragments in the actualscene For example when analyzing the position-sensitive

(a) (b)

Figure 1 Differences in movement behavior between (a) walking users and (b) driving users

c

ab

θ1

θ2

Corner

αα

α

Figure 2 Corner of userrsquos mobile behavior trajectory


trajectory fragments we can focus on the position vectors andthe weights WD WS WA 0 WL 1 are also feasible

According to the feature vector and its weight to completethe structural similarity comparison mainly through theanalysis of the differences between the feature vectors of thetrajectory segments to complete the comparison [14] accordingto the definition of the trajectory segment structure we candefine two trajectory segments are Li Lj 1le ine jle n ecomparison function of two trajectory segments is D(Li Lj)velocity vector is S(Li Lj) angle vector is A(Li Lj) and po-sition vector is L(Li Lj) e four comparison functions aboveconstitute the calculation of structural similarity of the trajectorysegments as shown in the following Formulas (7) and (8) efunction N(middot middot middot) denotes the normalization of the distanceBecause the range of each eigenvector in the trajectory segmentis different the normalization of the distance is the normali-zation of the distance of each eigenvectore SSIMof structuralsimilarity is represented by 1 minus the normalization of thedistance

S Li Lj1113872 1113873 D times WD + S times WS + A times WA + L times WL( 1113857

(7)

SSIM Li Lj1113872 1113873 1minusN S Li Lj1113872 11138731113872 1113873 (8)

e structural similarity comparison of trajectory frag-ments can express the structural differences of each tra-jectory fragment on the feature vectors erefore thesmaller the SSIM value of the trajectory fragmentsthe greater the SSIM value of the trajectory fragmentsMoreover the distance between the structural similaritiesof the trajectory fragments is symmetrical that isSSIM(Li Lj) SSIM(Lj Li) erefore it can be found thatthe method based on structural similarity can well reflect thestructural differences between trajectory segments

According to structural similarity the direction in-formation speed information angle information and po-sition information are compared [15]

(1) e direction vector comparison function D(Li Lj)

denotes the degree of similarity of two similar trajectorysegment Li Lj in the direction of motion As shown inFigure 3(a) φ is the angle between the direction of thetrajectory and the formula for calculating directionvector comparison function is as follows

D Li Lj1113872 1113873 Li

times sin ϕ if 0deg le 90deg( )

Lj

if 90deg le 180deg( )

⎧⎪⎨

⎪⎩(9)

If two similar trajectory fragments have the same di-rection and the angle φ is small the two trajectory fragmentstend to be parallel in the same direction which is called thebest state then the Dir Dist value approaches zero If twosimilar trajectory fragments are in opposite directions and thetwo trajectory fragments with larger angle φ tend to be inreverse parallel the worst condition is that the Dir Dist valueis the length of the trajectory fragments involved in thecomparison

(2) e speed vector comparison function S(Li Lj) ex-presses the trend of user mobility e velocity vectorcomparison function is shown in Formula (10) whereSmax(Li Lj) is |Vmax(Li)minusVmax(Lj)| representingthe absolute value of the maximum velocity differencebetween the trajectory segments SimilarlySavg(Li Lj) and Smin(Li Lj) represent the absolutevalue of the difference between the average velocityand the minimum velocity respectivelyWe can judgethe difference of velocity vectors from the three as-pects of maximum minimum and average velocity

S Li Lj1113872 1113873 13

Smax Li Lj1113872 1113873 + Savg Li Lj1113872 1113873 + Smin Li Lj1113872 11138731113872 1113873

(10)

(3) e angle vector comparison function A(Li Lj)

expresses the degree of eigenvalue change caused bythe change of direction in the trajectory segment Asshown in Formula (11) where the angle of rotationθ is calculated according to Formula (6) the in-ternal rotation angle is positive and the externalrotation angle is negative the angular distance ofthe trajectory segment is the cumulative value ofmany internal corners of the trajectory and thedirection of change within the trajectory segmentcan determine the value of each angle

A Li Lj1113872 1113873 1113936

P Li( )P Lj( 111385711 θi minus θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)

Figure 3(b) shows that if each corner of the two tra-jectory segments rotates to Li and Lj matches the value ofthe angle vector comparison function is 0 which is the bestcase If the two trajectory segments turn to Li and Lj inopposite directions that is the two trajectory segments are

Step 1 one by one scanning the location information point sequence in the user movement behavior trackStep 2 formula (5)Step 3 formula (6)Step 4 Set a threshold ω for corner θ store the corner satisfying |θ|gtω as a mutation point and then divide the track segmentaccording to the position information point of the corner n is the number of sampling points and the time complexity of thealgorithm is O(n)

ALGORITHM 1


in opposite jagged shape and the value of the angle vectorcomparison function is 1 this is the worst case

(4) For the position vector comparison functionL(Li Lj) we can use Hausdorff distance to measurethe location distance of the trajectory segment asshown in the following formula

L Li Lj1113872 1113873 max h Li Lj1113872 1113873 h Lj Li1113872 11138731113872 1113873 (12)

where h(Li Lj) maxaisinLi

(minbisinLj

(dist(a b))) is the direct

Hausdorff distance between Li and Lj ie the maxi-mum distance from a point in Li to the nearest Lj anddist(a b) represents the Euclidean distance functionbetween points

32 Similarity Computation of Userrsquos Mobile BehaviorTrajectory At present we collect and store the trajectoriesof tourism usersrsquo mobile behavior cluster the typicalsimilar trajectories from these trajectory data analyze thebehavior patterns of userrsquos mobile behavior trajectoriesand predict the personalized needs of tourism users basedon structural characteristics Clustering analysis is to divideuser behavior trajectory into several groups with highcohesion and low coupling It requires high similarity ofuser behavior trajectory in the same group and low sim-ilarity of user behavior trajectory in different groups egoal of clustering analysis is to find out the trajectory datawith the same or similar behavior patterns from the tra-jectories of some usersrsquo mobile behaviors analyze thepersonal preferences consumer demands and behaviorcharacteristics of the trajectories of tourism usersrsquo mobilebehaviors and accurately determine the similarity betweentrajectories of usersrsquo mobile behaviors At the same timethe trajectories of usersrsquo mobile behaviors with high sim-ilarity are gathered into one class [16]

Most of the online travel users are in the same scenicspot similar routes to carry out activities and most of theresulting mobile behavior trajectory data have localsimilarity and global dissimilarity It is difficult to find thepersonalized characteristics of tourism users by analyzingthe complex and large number of usersrsquo mobile behaviortrajectories and effectively extract users e analysis of apart of the mobile behavior trajectory is more conduciveto finding the information contained in it [17] ereforethe trajectory analysis method based on the whole tra-jectory in traditional research is easy to cause the inac-curacy of trajectory analysis In this paper we use

structural features to calculate the similarity of usermovement behavior is method needs to calculate everycorner of the userrsquos mobile behavior trajectory and findthe sampling point with larger rotation angle which isregarded as the sudden change point of the userrsquos mobilebehavior and then divides the trajectory segment by thesudden change point In this way the rotation angle ofeach trajectory segment obtained will not change sig-nificantly and the trajectory structure tends to be stableen a trajectory model of userrsquos mobile behavior isconstructed which is characterized by trajectory di-rection trajectory speed trajectory angle and trajectorydistance Taking these features as parameters thresholdvalues are set to express and adjust the weights of eachfeature according to the actual application scenarios anda trajectory similarity algorithm is constructed to calcu-late the userrsquos movement behavior e object of thispaper is to calculate the structural similarity of sometrajectory segments which are divided according to thesudden change points of large turning angles by using thetrajectory similarity algorithm constructed with structuralfeatures as parameters It is used to judge the similaritydegree of each userrsquos moving behavior trajectory and thencompletes the feature analysis of userrsquos moving behaviortrajectory e simulation results show that the trajectorysimilarity calculation algorithm is efficient the weightadjustment of each structural feature is flexible and thetrajectory analysis results are more in line with the needsof practical application scenarios and have higher ap-plication value and practical significance

On the basis of obtaining the feature vector distance ofuserrsquos moving behavior trajectory segment the trajectorysegment with high similarity is analyzed and then theclustering algorithm is used to complete the clustering ofuserrsquos moving behavior trajectory By comparing thestructural similarity between the trajectory segments andother trajectory segments which are not on the same tra-jectory a number of ε-nearest neighbor sets of trajectorysegments are formed e number of ε-nearest neighbor setsis used to determine the midpoint of trajectory segmentclustering and then the trajectory segment clustering isrealized A trajectory segment clustering algorithm based onstructural similarity is constructed

e steps of clustering algorithm based on structuralsimilarity are given in Algorithm 2

From the analysis of the above algorithms it can be seenthat in the trajectory segment clustering algorithm based onstructural similarity it is very important to determine the

Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)

Figure 3 Comparison of track direction and rotation angle (a) direction contrast (b) corner contrast


threshold value of ω ε-nearest neighbor and the thresholdvalue of σ nearest neighbor number which can directly affectthe time complexity and space complexity of the algorithmIt needs to be verified repeatedly and determined accordingto the actual application fields erefore we mainly analyzethe algorithm qualitatively

rough repeated verification of the algorithm in thedata analysis of trajectory of travel userrsquos movement be-havior the value of ω cannot be set too small and if set toosmall some characteristic details of trajectory segments willbe lost On the contrary the value ofω cannot be set too largeand cannot effectively identify the abrupt change point orsampling abnormality of the trajectory segment which di-rectly affects the structure of clustering analysis Similarly ifthe threshold value σ of the number of neighbors is set to belarge enough then no trajectory segment can satisfy therequirement of |Nε(L)|ge σ and all trajectory segments willbe marked as abnormal conditions On the contrary if σ isset too small all the trajectory segments may become theclustering center so that the trajectory segments will be self-contained and the number of clusters will be too large

33 Discovery of Popular Tourist Attractions By effectivelyidentifying the location information points in the trajectorydata of usersrsquo mobile behavior the feature vectors of thetrajectory segments can be extracted and the semantics ofthese location information points can be expressed as theroute the scenic spots and the behavior patterns of anonline travel user in the past period of time By clusteringand analyzing the trajectory fragments containing loca-tion information points we can find that the travelingusers have a longer time in a certain area which can beinterpreted as the tourist users have a higher degree ofinterest in a certain scenic spot Semantic expression is apopular tourist spot with longer stay time for online travelusers In practical scenarios many traveling users will visitthe same or similar scenic spots From the trajectory ofusersrsquo mobile behavior and the region of interest travelingusers with similar trajectory and the same region of in-terest can predict their similar preferences or similarbehavior characteristics ese regions of interest fre-quently stayed by tourist users will appear as overlappingregions in the trajectory of userrsquos mobile behavior If these

overlapping regions are found the popular scenic spotsconcerned by tourist users can be found and the users wholike these scenic spots can be clustered And then dig outthe other characteristics of these users to complete thepersonalized tourist attractions recommendation ofsimilar tourists We extract the feature parameters of theseoverlapping areas such as overlap time and overlap timeswhich can reflect the similarity between the traveling users Itcan identify the tourist attractions that the tourist users areinterested in during the mobile process and recommend themost likely popular tourist inventory for other tourist userswho have a higher similarity with their userrsquos mobile behaviortrajectory so as to tap the potential preferences of the touristusers [18] Assuming that travel user A and travel user B sharea higher degree of similarity in the trajectory of usersrsquo mobilebehavior it can be found that some scenic spots are visited byusers A but not by users B rough mining it is known thatthese scenic spots may be of interest to users B en we canrecommend these scenic spots to B users through A users sothat these scenic spots become the potential and most likelyscenic spots for B users to visit We can also use the activitysequence to express the popular scenic spots that touristsoften visit and the trajectory of nearby tourists is more in-structive [19]

In the process of analyzing mobile user behavior tra-jectory data with structured eigenvectors it is not difficult tofind that the moving speed of user behavior trajectory is notthe same in different time periods or it is slower in a certaintime period or it is faster in a certain time period Figure 4shows the moving speed of the user in different periods oftime in which the trough is formed during the period whenthe user moves slowly and the peak is formed during theperiod when the user moves fast but both trough and peakcan indicate the userrsquos continuous generating activity Andthe slow moving trough time contains more user behaviorcharacteristics so this paper focuses on the behaviorcharacteristics of mobile users in trough situation

As shown in Figure 4 by comparing the speed distanceand time of userrsquos moving behavior trajectory the structuredfeatures of mobile users and the behavior differences oftourist users can be clearly analyzed We focus on theanalysis of two dimensions the speed and the time of theslower wave trough As shown in Figure 5 the slower thetraveling speed of the tourist user or the less the change of

Step 1 first calculate the corner θ of each track segment sampling point PiStep 2 according to the corner threshold ω we divide the trajectory of user movement into TS of some track segmentsStep 3 calculate the distance between the trajectory feature vectors based on the weight of the trajectory segment feature vectorsStep 4 calculate the ε-nearest neighbor set of the track segments with high similarityStep 5 the distance clustering segment is centred on the similarity track segment ε-nearest neighbor setStep 6 initialize clustering ID and track segment clustering markersStep 7 traverse the trajectory fragments find the core clustering and set the clustering ID and then add the pointers of thesetrajectory fragments to a new node in the index treeStep 8 determine whether the set center of ε-nearest neighbors meets the specified distance If it meets the requirement then add thecluster ID marker to the trajectory fragment expand the clustering construct the index tree node and repeat steps 7 and 8 until alltrajectory fragments are traversed

ALGORITHM 2


the active area in a period of time the most likely thepredictable user behavior characteristic is that is thetraveling user stays at a certain scenic spot for browsingresting or taking photos e longer the trough the moreattractive the scenic spot is e more tourists are staying atthe same scenic spot themore scenic spots can be designatedas popular tourist attractions

For example when a tourist visits a scenic spot he or sheforms a trajectory of the userrsquos movement behavior reetroughs appear in the trajectory indicating that the user mayhave experienced three scenic spots or rest areas of whichthe first trough has a shorter experience It shows thattourists spend less time visiting the first scenic spot travelfaster continue to move forward at a faster speed after thetour and spend a little more time watching the tide or takingpictures when they meet the scenic spot of interest Sotourists will slow down move in a more fixed area and travelat a slower speed thus appearing the second trough periodafter the tourists continue to move forward when theformation of the third trajectory speed reached trough statesemantic expressionmay have two situationse first is thatthe tourists reach a certain degree of fatigue or meet a restarea stop and rest the second is that the tourists arrive at awell-known scenic spot gather more tourists people willstay in a certain position waiting for sightseeing andphotography moving slowly and almost stop e abovetwo semantics can be distinguished by whether the locationin the electronic map is a resting area or a scenic spotHowever in the actual tourist attractions the situation maybe more complicated For example a tourist is an outdoorsports enthusiast who has good physical strength and likesnatural scenery Because of his fast moving speed there islittle difference between the wave crest and trough of thewaveform trajectory formed by the speed and distance

Although his tour speed is fast and his stay time is short thelocation he stays in is still the area of interest In this waymoving objects with similar frequencies in the velocity-distance waveform can be found not only in the knownhot spots of the users but also in the scenic spots that thepotential users may be interested in even in the preferencesoccupations and personality characteristics of the touristusers It helps to gather tourists with similar preferences andsimilar personalities to achieve the confluence module [20]

Popular scenic spots refer to scenic spots with longstaying time after arrival [21] In the userrsquos mobile behaviortrajectory the hot spots can be marked as H

H1 H2 Hn1113864 1113865 Hj Li Li+1 Lm1113864 1113865 H is used to de-note a trajectory fragment When the traveling user passesthrough a hot spot area with high interest and stays for a longtime the trajectory fragment moves at a speed close to or farbelow the normal trajectory speed We can think that thetourist user has conducted a deep browsing in the scenic spotor some behavior activities have taken place in the scenicspot area We can analyze the information such as the timeof arrival the time of stay and so on e region with denseuser access points can be expressed as a popular touristattraction area with high user access frequency [22]

Because GPS receiving equipment receives satellite signalsin vast and open areas with high intensity and good posi-tioning effect satellite signals in indoor areas will be shieldedby the wall resulting in weak positioning signal and reducedpositioning accuracy [23]erefore when analyzing touristsrsquopreference for scenic spots through the status of stay it isnecessary to distinguish between outdoor and indoor scenicspots e positioning signal of outdoor scenic spots is goodand has high precision It can acquire the location in-formation points at sampling frequency in real time and formthe locus of userrsquos movement with dense location informationpoints e positioning signal of indoor scenic spots is weakwhich affects the positioning accuracy Even when the signalis lost the location information points cannot be obtained intime according to the sampling frequency requirement andthe space area of the indoor attractions is small which makessome location information points overlap is repetitiveactivity can also find that the tourists are visiting a certainindoor attractions regularly e popular scenic spots aredivided into two types one is the outdoor scenic spots such asnatural landscape gardens playgrounds and other broadareas in a longer period of time can obtain more denselocation information points formed by the userrsquos mobilebehavior trajectory recorded as HRII Another kind is indoorscenic spots such as restaurants shopping malls touristcenters and other closed areas in a long period of time maylose a certain location information point sampling in-formation but after leaving the area they can get the locationinformation point again recorded as HRI Firstly the tra-jectory data of userrsquos mobile behavior are obtained by sam-pling the location information points and then the trajectorydata of userrsquos mobile behavior is denoised Finally accordingto the characteristics of the location information points theHRI and HRII popular scenic spots domain are divided by thedensity clustering method e steps for finding popularscenic spots is given in Algorithm 3

Spee

d

Distance

Figure 5 Two-dimensional trajectory analysis of user movementbehavior

Speed

DistanceTime

Figure 4 Mobile speed of user behavior trajectory


Because the location information points sampled byGPS are affected by the factors of region space andweather it is easy to have inaccurate positioning or in-terruption of positioning [24] When the user enters theindoor scenic spots from the outdoor scenic spots the shortsignal interruption will occur and the positioning datareceiving error will easily occur In order to adapt thiserror a maximum disturbance threshold MT is set in thehot spot detection algorithm to enhance the accuracy oflocation information points

e outdoor and indoor scenic spots are divided accordingto the location information of popular scenic spotse density-based hot spot discovery algorithm is used to retrieve twodifferent types of user residence areas outdoor and indoor inthe trajectory of userrsquos mobile behavior and define them as hotspots [25] e algorithm has four input parameters trajectoryof userrsquosmobile behaviorminimum speedminimum time andmaximum disturbance threshold Among them the thresholdofminimum speed is related to the activity speed of tourist usersin the scenic spot area If walking tour the general speed is 2 to3meters per second If the sudden reaction speed of the locationinformation points slows down significantly it indicates that thetourist users have arrived at the scenic spot area and specificuser behavior has taken place On the contrary if the suddenresponse speed of the location information points increases overa period of time it indicates that the travel users have changedtheir behavior patterns For example we can leave the scenicspot and take a sightseeing bus to the next scenic spot so we can

set it according to the sampling timeere is no absolute fixedvalue in the setting of the minimum time thresholdGenerally if a tourist user stays in a certain area for morethan 30minutes it can be considered that the tourist userarrives at a scenic spot or rest area and changes in userbehavior have taken place resulting in a specific activitye maximum perturbation threshold is only used to ex-press the number of continuous perturbations when theabnormal location information points are sampled If thenumber of abnormal location information points is smallerthan the perturbation threshold the abnormal samplinginformation points can be merged into the normal locationinformation points of the userrsquos mobile behavior trajectorydata set If the number of abnormal location informationpoints sampled in the former HR is larger than the per-turbation threshold it is necessary to preserve the currentuserrsquos mobile behavior trajectory and then detect the newhot spots after abnormal location information pointssampled to form a new userrsquos mobile behavior trajectorye settings of the minimum time threshold and themaximum disturbance threshold are related to the sam-pling frequency of the location information points In thepopular scenic spot area detection algorithm the trajectoryof the userrsquos movement behavior is traversed twice ealgorithm complexity is linear order O(n) Among them ndenotes the number of location information points in thetrajectory of the userrsquos mobile behavior and the algorithmcan retrieve the hot spots with frequent activities

Input parameters user movement behavior trajectory Q minimum speed S minimum time T and maximum disturbance thresholdMTOutput parameters collection of popular scenic spots HRStep 1 for (i 2 ile |Q| i++) lowast | Q| represents the number of location information pointslowastStep 2 D [iminus 1] middotTcal T (piminus 1 pi)Step 3 D [iminus 1] middot S cal D (piminus 1 pi)D [iminus 1] middotTStep 4 HR C CO falseStep 5 for (j 2 jle |T|minus 1 j++) lowastcycle search indoor attractions area HRIlowastStep 6 if (D [jminus 1] middotTgtT and D [iminus 1] middot Slt nlowast S) thenStep 7 C pjminus 1 pj lowastrecord location information point stay arealowastStep 8 if (not CO) then CO trueStep 9 else CUpdate pjminus 1 pj lowastmerge the location information points closer to the collection HRlowastStep 10 else if (CO) thenStep 11 HR C CO false C lowastsearch outdoor attractions area HRIIlowastStep 12 if (D [jminus 1] middot Sle S) then lowastdetermine activity intensive areaslowastStep 13 C pjStep 14 if (not CO) then CO trueStep 15 else if (CO) thenStep 16 last Index look Ahead (MT S)Step 17 if (last Indexle j+MT) thenStep 18 for k last Index downto j doStep 19 C pkStep 20 j last IndexStep 21 else if (time (C)gtT) thenStep 22 HE C C CO falseStep 23 return HR

ALGORITHM 3


4 Travel Recommendation Model

41 Application Scene

(1) Tourist Attractions Recommendation In the mobilescenario where the user completes the tourist at-traction tour the tourist user will generate multi-dimensional information to realize the applicationscenario recommended by the tourist attractionwhich can hide the rarely used dimensions Focus onthe userrsquos city mobile behavior trajectory in thecorner location information points hot spots anduser behavior patterns and other dimensions ac-curate analysis of user preferences the same char-acteristics of the user recommend the most likelyfavorite tourist attractions [26]

(2) Hotel Catering-Related Recommendation Touristusers need to visit other applications frequently in theprocess of touring such as electronic map applica-tions O2O applications restaurants e-commerceoutdoor equipment etc a single visit to each ap-plication users need to frequently exit one applicationlogin another application will cause user in-convenience inefficiency etc According to the ap-plication relevance assessment model we analyze theapplication of tourism usersrsquo preferences Establishthe relationship between these applications so thatusers can access a travel APP directly related to otherapplications they are used to daily life making travelAPP a personalized all-round service platform [27]

(3) Tourism User Preference Content RecommendationTourist users may often inquire about certain contentssuch as outdoor equipment fitness and health carerestaurants and entertainment and surrounding scenicspots in the process of using the application Accordingto the content retrieval evaluation model users canretrieve keywords learn user preference content anddiscover their hobby characteristics and behaviorcharacteristics According to these preferences travelAPP can construct a personalized interface for usersgiving priority to the information articles and newsthat travel users are interested in [28]

42 Travel Recommendation System With the combinationof Bluetooth WIFI and other RF communication tech-nologies and mobile terminal devices mobile point-to-pointcommunication environment has been derived and manydifferent research topics have also emerged is researchtalks about the relationship between mobile attractionrecommendation system and social software from theperspective of mobile social software and completes in-teraction through user comment sharing in mobile point-to-point environment In this paper an interactive system oftourism comment information sharing and social net-working software is established which includes threefunctions recommendation reunion and comment It isused to explore the interaction between users in mobilepoint-to-point environment e preliminary test has been

completed in this paper e experimental results show thatthe recommendation convergence and comment functionsof the system can provide precise services for users andprovide a basis for further research on the wide applicationof user behavior trajectory in precise marketing

is paper focuses on the problem of informationsharing and social interaction of tourism mobile recom-mendation system in mobile point-to-point environmente system mainly includes three functions recommen-dation convergence and information sharing In the rec-ommendation function section we assume that users willleave comments and other information after visiting a scenicspot When other users meet with them they can exchangecomments through RF communication technology esecomments are calculated by system algorithm to recom-mend scenic spots that meet usersrsquo interests In additionusers can also actively send requests to other people to jointhe information and find similar interests around users tovisit a scenic spot Of course users can also actively sharelocation comments traffic conditions tourist density andother related information

In order to enable the interaction and sharing of in-formation between remote users the relay mode undermobile point-to-point can be adopted Every user in thesystem plays the role of information transmission that is tosay each userrsquos mobile terminal is a relay node for in-formation transmission and constantly transfers the in-formation they have mastered to the users at a long distance

421 System Architecture e mobile peer-to-peer envi-ronment mainly transmits information through the directtransmission between peer-to-peer users and the relay modeassisted by the third party Using this feature the systemproposed in this paper mainly provides three servicesrecommendation convergence and review First of all themain purpose of recommendation service is to recommendscenic spots similar to userrsquos interests to users through userrsquosevaluation information so that users can have a referencedirection for the next destination in the journey so thatusers can travel more smoothly Secondly the convergenceservice allows users to initiate a convening activity gatherother interested users around visit the scenic spots togetheror buy specialty goods together through group buying to geta better price or to strive for preferential services irdlyevaluation services are divided into general information andspecific information General information is simply thetransmission of personal information and specific in-formation so that the use of convergence services conveys theconvening activities of the department of the offensivethrough short messages and the expression of personal in-formation is incompatible specific information is only forconvening activities issued information To provide the aboveservices the system architecture is presented in Figure 6

(1) Interface Module is module is responsible for the userand the system function docking through this module thesystem function interface is expressed and the user is guidedto carry out the operation of various functions


(2) Scoring Module At present the commonly used rec-ommendation system is based on the scoring mechanismwhich collects userrsquos scoring data to calculate and providerecommendation services e scoring module proposed inthis study is mainly responsible for recording the userrsquosevaluation of scenic spots At the initial stage the recom-mendation system often faces the problems of incompleteuser scoring information too many items not scored and thedifficulty of calculation caused by the noise of data resultingin the decline of recommendation accuracy erefore thisstudy classifies scenic spots requires users in the initial stagemust be based on the type of scenic spots scoring to ensurethat individual users in the initial stage and needs to score forthe type of scenic spots to ensure that individual usersrsquo scoringinformation has been scored by the common column

(3) Transport Module Because this research system is built inthe mobile point-to-point environment user scoring in-formation and other need to be obtained and transmittedthrough the transmission function this study uses Bluetoothtransmission technology to achieve the transmission of re-lated functions is module enables the system to auto-matically exchange scoring data through Bluetooth withoutinterfering with the user when they meet so as to achieve thepurpose of collecting data In terms of scoring exchangemechanism this study currently uses unlimited scoring ex-change method when users meet the exchange of all thescoring data held by both sides However information thathas not yet been scored is not helpful to the recommendationsystem erefore in the scoring exchange module it is as-sumed that only the user has scored more than five scenicspots before the exchange while the other scoring data

obtained by others is more than five items before passing onto other users On the other hand the transmission modulehas the search function and can discover other users aroundthe user when the user wants to pass its ideas to the sur-rounding users it can be completed through this module

(4) Recommendation Module is study analyzes the rec-ommended operation by exchanging accumulated scoredata is recommendation module uses collaborative rec-ommendation and Pearson correlation coefficients to per-form recommendation operation e formula is shown in(13) Suppose that U(i a) is used to predict the possibledegree of preference of i to a scenic spots Fj(a) is the scoreof user j for a attractions and Fi is used to score the averagescore of holder i Fj is the average score of user j andsim(i j) is the similarity between user i and user j calculatedby Pearson correlation coefficient

U(i a) Fi +1113936jsim(i j) times Fj(a)minusFj

11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)

In the process of recommendation the recommendationmodule first calculates Pearson correlation coefficient cal-culates the first 20 items of scoring data which aremost similarto users and then runs the subsequent recommendation al-gorithm Finally the userrsquos predicted value of a certain scenicspot is obtained by weighted average of these scoring data andsimilarity and five scenic spots with the highest predictedvalue are recommended to users for reference

(5) Position Module is module can use Bluetooth GPSreceiver to receive satellite signals and select the local latitude

Current user

Interface module

Scoringmodule

Privacymodule

Recomme-ndationmodule

Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users

Figure 6 Mobile point-to-point tourism recommendation system architecture


and longitude values to determine the location of the userFinally combined with the processing results of the recom-mendation module the electronic map shows the location ofeach recommendation site and the location of the user

(6) Information Module e concept of userrsquos active com-ment can be added in the system through the transmissionof information users can express their personal ideas toother users around Adding the function of transmittinginformation in this part the user can not only transmit thenew information but also transmit the received informationto other users in the transmission range

(7) Convergence Module Since the system is designed in amobile peer-to-peer environment a mobile convergencefunction is derived from the concept of mobile social net-works rough this function travelers can dynamicallysearch for other users with the same goals and preferencesrough the transmission of information travelers can sharethe requested information to the surrounding users and thusfind travelers willing to act together

(8) Privacy Module One of the focuses of mobile socialsoftware is to explore the interaction between users but noteveryone is willing to interact with others so this study addsprivacy considerations is module can provide userswhether to allow all other users or only allow some friends tosearch their own location through the system through theprivacy settings users can not be disturbed by other users tocarry out system operations but also to observe whetherthere is a willingness to interact between users

J2ME can be chosen as the development platform of thesystem and Bluetooth technology is the basic wirelesstransmission technology commonly available in mobileterminals erefore it is feasible to use Bluetooth tech-nology as a transmission tool In order to expand the scopeof information transmissionWIFI wireless network can alsobe considered as a transmission medium which can effec-tively solve the problem of short transmission distance andunstable signal of Bluetooth

5 Conclusion

Advanced GPS devices enable people to record their locationhistories with GPS trajectories e trajectory of usersrsquomobile behavior means to a certain extent that a personrsquosbehavior and interests are related to their outdoor activitiesso we can understand the users and their locations and theircorrelation according to these trajectories is informationenables accurate travel recommendations and helps peopleto understand a strange city efficiently and with high qualityBy measuring the similarity of different user location his-tories the similarity between users can be estimated andpersonalized friend recommendation can be realized euser stereoscopic user portrait can be portrayed through theintegration of user movement behavior trajectory and socialinformation is paper takes the trajectory data of tourismusersrsquo mobile behavior as the research object and constructsthe tourism precise marketing model In the process of

obtaining the trajectory of user movement the character-istics of mobile user behavior track data are taken intoaccount e sensitivity of various features in the trajectoryanalysis process is adjusted by weighte structured featurevectors and popular scenic spots discovery methods of userrsquosmobile behavior trajectory are fully studied by clustering andcollaborative filtering techniques which lay a foundation forconstructing the application model of tourism precisionmarketing

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

is research work was supported by the Ministry of Ed-ucation Humanities and Social Sciences Planning FundProject (No 18YJAZH128) and Research Project of HarbinUniversity of Commerce (No 18XN022)

References

[1] Y Yuan and M Raubal ldquoMeasuring similarity of mobilephone user trajectoriesndasha Spatio-temporal Edit Distancemethodrdquo International Journal of Geographical InformationScience vol 28 no 3 pp 496ndash520 2014

[2] Z Sun and X (Jeff) Ban ldquoVehicle classification using GPSdatardquo Transportation Research Part C Emerging Technologiesvol 37 pp 102ndash117 2013

[3] D Wang ldquoApproaches for transportation mode detection onmobile devicesrdquo in Proceedings of Seminar on Topics in SignalProcessing pp 77ndash82 2014

[4] S Hong and A Vonderohe ldquoUncertainty and sensitivityassessments of GPS and GIS integrated applications fortransportationrdquo Sensors vol 14 no 2 pp 2683ndash2702 2014

[5] M Lin and W-J Hsu ldquoMining GPS data for mobility pat-terns a surveyrdquo Pervasive and Mobile Computing vol 12pp 1ndash6 2014

[6] S Khajehzadeh H Oppewal and D Tojib ldquoMobile couponswhat to offer to whom and whererdquo European Journal ofMarketing vol 49 no 5-6 pp 851ndash873 2015

[7] H Li ldquoReview on state-of-the-art technologies and algorithmson recommendation systemrdquo in Proceedings of the In-ternational Conference on Mechatronics Engineering and In-formation Technology (ICMEIT 2016) p 7 Wuhan ZhichengTimes Cultural Development Co Xirsquoan China 2016

[8] Htet Htet Hlaing ldquoLocation-based recommender system formobile devices on University campusrdquo in Proceedings of 2015International Conference on Future Computational Technol-ogies (ICFCTrsquo2015) International Conference on Advances inChemical Biological amp Environmental Engineering (ACBEE)and International Conference on Urban Planning Transportand Construction Engineering (ICUPTCErsquo15) p 7 UniversalResearchers in Science and Technology Universal Re-searchers in Science and Technology Singapore March 2015


[9] W Worndl and B Lamche ldquoUser interaction with context-aware recommender systems on Smartphonesrdquo icom vol 14no 1 2015

[10] L O Colombo-Mendoza R Valencia-Garcıa G Alor-Hernandez and P Bellavista ldquoSpecial issue on context-awaremobile recommender systemsrdquo Pervasive and Mobile Com-puting vol 38 pp 444-445 2017

[11] L O Colombo-Mendoza R Valencia-Garcıa A Rodrıguez-Gonzalez G Alor-Hernandez and J J Samper-ZapaterldquoRecomMetz a context-aware knowledge-based mobile rec-ommender system for movie showtimesrdquo Expert Systems withApplications vol 42 no 3 pp 1202ndash1222 2015

[12] W-S Yang and S-Y H iTravel ldquoA recommender system inmobile peer-to-peer environmentrdquo Journal of Systems ampSoftware vol 86 no 1 pp 12ndash20 2013

[13] T Pessemier D Simon K Vanhecke B Matte E Meyns andL Martens Context and Activity Recognition for PersonalizedMobile Recommendations Springer Berlin Germany 2014

[14] J Zeng F Li Y Li J Wen and Y Wu ldquoExploring the in-fluence of contexts for mobile recommendationrdquo In-ternational Journal of Web Services Research vol 14 no 4pp 33ndash49 2017

[15] A Majid L Chen G Chen H T Mirza I Hussain andJ Woodward ldquoA context-aware personalized travel recom-mendation system based on geotagged social media dataminingrdquo International Journal of Geographical InformationScience vol 27 no 4 pp 662ndash684 2013

[16] D Gavalas C Konstantopoulos K Mastakas andG Pantziou ldquoMobile recommender systems in tourismrdquoJournal of Network and Computer Applications vol 39pp 319ndash333 2014

[17] S K Hui J J Inman Y Huang and J Suher ldquoe effect of in-store travel distance on unplanned spending applications tomobile promotion strategiesrdquo Journal of Marketing vol 77no 2 pp 1ndash16 2013

[18] L Liu J Xu S S Liao and H Chen ldquoA real-time personalizedroute recommendation system for self-drive tourists based onvehicle to vehicle communicationrdquo Expert Systems withApplications vol 41 no 7 pp 3409ndash3417 2014

[19] J P Lucas N Luz M N Moreno R Anacleto A AlmeidaFigueiredo and C Martins ldquoA hybrid recommendationapproach for a tourism systemrdquo Expert Systems with Appli-cations vol 40 no 9 pp 3532ndash3550 2013

[20] Z Bahramian R Ali Abbaspour and C Claramunt ldquoACONTEXT-AWARE tourism recommender system basedON a spreading activation methodrdquo ISPRSndashInternationalArchives of the Photogrammetry Remote Sensing and SpatialInformation Sciences vol XLII-4W4 pp 333ndash339 2017

[21] M Nilashi K Bagherifard M Rahmani and V Rafe ldquoArecommender system for tourism industry using cluster en-semble and prediction machine learning techniquesrdquo Com-puters amp Industrial Engineering vol 109 pp 357ndash368 2017

[22] I Cenamor T de la Rosa S Nuntildeez and D Borrajo ldquoPlanningfor tourism routes using social networksrdquo Expert Systems withApplications vol 69 pp 1ndash9 2017

[23] K Meehan T Lunney K Curran and A McCaugheyldquoAggregating social media data with temporal and environ-mental context for recommendation in a mobile tour guidesystemrdquo Journal of Hospitality and Tourism Technology vol 7no 3 pp 281ndash299 2016

[24] Z Shi and A B Whinston ldquoNetwork structure and obser-vational learning evidence from a location-based socialnetworkrdquo Journal of Management Information Systemsvol 30 no 2 pp 185ndash212 2014

[25] Q Lu ldquoMobile e-commerce precision marketing model andstrategy based on LBSrdquo E-Business Journal no 4 pp 20-212014

[26] P J Danaher M S Smith K Ranasinghe and T S DanaherldquoWhere when and how long factors that influence the re-demption of mobile phone couponsrdquo Journal of MarketingResearch vol 52 no 5 pp 710ndash725 2015

[27] N M Fong Z Fang and X Luo ldquoGeo-conquesting com-petitive locational targeting of mobile promotionsrdquo Journal ofMarketing Research vol 52 no 5 pp 726ndash735 2015

[28] S A Shad and E Chen ldquoPrecise location acquisition ofmobility data using cell-idrdquo International Journal of ComputerScience Issues vol 9 no 3 2012


Computer Games Technology

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

Advances in

FuzzySystems


Volume 2018


ReconfigurableComputing



Applied Computational Intelligence and Soft Computing

thinspAdvancesthinspinthinsp

thinspArtificial Intelligence

Hindawiwwwhindawicom Volumethinsp2018


Civil EngineeringAdvances in


Electrical and Computer Engineering

Journal of

Journal of

Computer Networks and Communications


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia


Biomedical Imaging



Engineering Mathematics


RoboticsJournal of



Computational Intelligence and Neuroscience


Mathematical Problems in Engineering

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018


Human-ComputerInteraction

Advances in


Scientic Programming

Submit your manuscripts atwwwhindawicom

According to the Statistical Bulletin on National Eco-nomic and Social Development of 2016 issued by the StateStatistical Bureau in the whole year of 2016 the number ofdomestic touristsrsquo trips reached 44 billion an increase of112 over the previous year and the income of domestictourism increased by 152 to 39390 billion yuan enumber of inbound tourists reached 13844 million anincrease of 35 and international tourism revenue in-creased by 56 to $120 billion e number of domesticresidents in China has reached 135 million 130 thousand anincrease of 57 [2] With the continuous promotion of thestrategic pace of building a well-off society in an all-roundway tourism has become an important part of the peoplersquosdaily life in China marking that Chinarsquos tourism industryhas entered the era of mass tourism

Based on this background in the mobile e-commerceenvironment based on LBS location service research andanalysis of userrsquos mobile behavior trajectory can extractvaluable userrsquos mobile behavior features from a largenumber of mixed dynamic data and integrate the mobilebehavior and consumer behavior of tourism users Basedon LBS location services will integrate the mobile be-havior and consumer behavior of tourism users thenexcavate the marketing value of consumers and timelyachieve the marketing objectives of enterprises on theappropriate media so that mobile e-commerce marketingbecomes more accurate and effective rough the re-search of this subject the interests of enterprises con-sumers and media can be maximized at the same timeproviding personalized products and services for mobilee-commerce improving consumer loyalty and corecompetitiveness of mobile e-commerce and bringinghigher profits for e-commerce enterprises [3]

12 Presentation of Problems e data of mobile terminalusersrsquo historical consumption behavior and locationmovement process are recorded and stored according tothe time series forming the userrsquos mobile behavior tra-jectory data which can be collected by multiple deviceterminals e userrsquos mobile behavior trajectory datacontain a lot of useful information Mobile behavior tra-jectory can express the behavior activities of mobile users inthe real world ese activities imply userrsquos interestshobbies experiences and behavior patterns [4] For ex-ample a userrsquos activities in a week may start from home towork every day and a user may go to shopping malls parksand other places on weekends erefore how to effectivelyutilize the userrsquos mobile behavior trajectory and extractuseful information from the userrsquos mobile behavior tra-jectory data is very important for the realization of per-sonalized recommendation service

From the view of consulting a large number of docu-ments there are more papers on location services than onmobile marketing However most of the previous articles onlocation services focused on the application of natural sciencesuch as surveying and mapping technology network devel-opment geographic information and so on In recent yearsthe number of cross-research articles onmanagement science

medicine and agriculture combined with location-basedservices has begun to increase most of which are the com-bination of location-based services and related industries tostudy the application or technology development of specificindustries For example the combination of location servicesand logistics technology can track the journey of packagesCombining location services with electronic maps can pro-vide catering entertainment discounts and other in-formation within a certain range according to the location ofusers Combining location service with utility technology canquickly find information such as tap water gas explosion andso on e main keywords of literature research include lo-cation service technology location service system locationservice terminal location service strategy mobile locationservice and so on Based on location the services industry iscurrently considered one of the most dynamic industriesWith the rapid development of mobile Internet and Internetof ings technology more debris time has been transferredto mobile phones tablets and smart products [5]

e characteristics of mobile marketing such as pre-cision interaction novelty and effective delivery are moreand more concerned and recognized by various industriesis paper focuses on the main characteristics of the end-users of the tourism industry such as frequent locationmovement strong sense of sharing and rich demand forservices First of all the users of online travel must bemobile users who usually do not stay in a location for a longtime and a high probability of frequent location changeswill produce a large number of location data Moreover ingeneral traveling users arrive at an unknown location ormove in a series of unfamiliar geographical environmentswhich makes travel usersrsquo demand for location-basedservices take precedence over personal privacy pro-tection and enable them to obtain real-time user locationinformation ese location data provide favorable con-ditions for our research Secondly the behavior of touristusers is quite different from that of ordinary people Inbeautiful scenery and not very familiar environment userswill spontaneously produce self-awareness Most peopleshare location photos and moods through social platformsandmicromessaging and travel companies can access thesesocial data to accurately portray users and provide accurateservices for them ird travel users need high-qualityservices to obtain high-quality tourism experience Sce-nic spots accommodation restaurants transportation andfinance including tour guides and their fellow travelers arealso important factors in achieving a high-quality tourismexperience ese rich demands for services will generateenormous commercial value [6] erefore this articleadopts top-down overall analysis to design ideas frombottom to top By analyzing the userrsquos characteristicsthrough the trajectory of userrsquos mobile behavior this paperconstructs a travel recommendation system in the mobilepoint-to-point environment and a precise marketing modelin the tourism industry based on the trajectory of userrsquosmobile behavior so as to provide appropriate services forthe appropriate users at the right time and place in order toprovide reference for relevant tourism enterprises toachieve precise marketing











1113868111386811138681113868 (1)






(2)











(4)

















2ab (5)







(a) (b)


c

ab

θ1

θ2

Corner

αα

α






(7)






D Li Lj1113872 1113873 Li


Lj


⎧⎪⎨

⎪⎩(9)



S Li Lj1113872 1113873 13


(10)



A Li Lj1113872 1113873 1113936


11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)



ALGORITHM 1






(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in













1113868111386811138681113868 (1)






(2)











(4)

















2ab (5)







(a) (b)


c

ab

θ1

θ2

Corner

αα

α






(7)






D Li Lj1113872 1113873 Li


Lj


⎧⎪⎨

⎪⎩(9)



S Li Lj1113872 1113873 13


(10)



A Li Lj1113872 1113873 1113936


11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)



ALGORITHM 1






(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in











(4)

















2ab (5)







(a) (b)


c

ab

θ1

θ2

Corner

αα

α






(7)






D Li Lj1113872 1113873 Li


Lj


⎧⎪⎨

⎪⎩(9)



S Li Lj1113872 1113873 13


(10)



A Li Lj1113872 1113873 1113936


11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)



ALGORITHM 1






(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in









2ab (5)







(a) (b)


c

ab

θ1

θ2

Corner

αα

α






(7)






D Li Lj1113872 1113873 Li


Lj


⎧⎪⎨

⎪⎩(9)



S Li Lj1113872 1113873 13


(10)



A Li Lj1113872 1113873 1113936


11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)



ALGORITHM 1






(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in







(7)






D Li Lj1113872 1113873 Li


Lj


⎧⎪⎨

⎪⎩(9)



S Li Lj1113872 1113873 13


(10)



A Li Lj1113872 1113873 1113936


11138681113868111386811138681113868

111386811138681113868111386811138681113874 1113875 θi

11138681113868111386811138681113868111386811138681113868 + θj

11138681113868111386811138681113868

111386811138681113868111386811138681113874 11138751113874 1113875

P Li( 1113857 + P Lj1113872 1113873 (11)



ALGORITHM 1






(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in








(minbisinLj









Sj

Si Li

Lj

ej

ei

Dir Dist

ϕ

(a)

Sj

Si

Li Ljejei

(b)










ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in











ALGORITHM 2








Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in










Spee

d

Distance


Speed

DistanceTime







ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in








ALGORITHM 3



















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in





















11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in









11138681113868111386811138681113868

11138681113868111386811138681113868

1113936jsim(i j) (13)



Current user

Interface module

Scoringmodule

Privacymodule


Positionmodule

Informationmodule

Confluencemodule

Database

Interface module

Other users








5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in









5 Conclusion



Data Availability




Acknowledgments


References




































Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in






























Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in









Advances in

FuzzySystems


Volume 2018













Journal of

Journal of



Hindawi


Advances in

Multimedia


Biomedical Imaging





RoboticsJournal of









Volume 2018



Advances in




Documents

Research on Precision Marketing Model of Tourism Industry