Intelligent Information Systems

Evolutionary

Step

Business Question Enabling

Technologies

Product

Providers

Characteristics

Data

Collection

(1960s)

"What was my total

revenue in the last

five years?"

Computers, tapes

, disks

IBM, CDC Retrospective,

static data

delivery

Data Access

(1980s)

"What were unit

sales in New England

last March?"

Relational

databases

(RDBMS), Struct

ured Query

Language

(SQL), ODBC

Oracle, Sybas

e, Informix, I

BM, Microsoft

Retrospective,

dynamic data

delivery at

record level

Data

Warehousing

&

Decision

Support

(1990s)

"What were unit

sales in New England

last March? Drill

down to Boston."

On-line analytic

processing

(OLAP),

multidimensional

databases, data

warehouses

Pilot,

Comshare,

Arbor,

Cognos,

Microstrategy

Retrospective,

dynamic data

delivery at

multiple levels

Data Mining

(Emerging

Today)

"What’s likely to

happen to Boston

unit sales next

month? Why?"

Advanced

algorithms,

multiprocessor

computers,

massive

databases

Pilot,

Lockheed,

IBM, SGI,

numerous

start-ups

(nascent

industry)

Prospective,

proactive

information

delivery

• Standard database operations presentresults to the user as they existed indatabases• A report showing the breakdown of salesby product line and region isstraightforward for the user to understandbecause they intuitively know that this kind

of information already exists in the database

Business Intelligence (BI) tools such as query and reporting are used to answer questions by the user

These questions deal primarily with the analysis of historical results and trends

- what were our sales in the past month in a certain region? - what were our most profitable products? - which of our suppliers were most reliable? - which customers generated the most revenue?

• Extracts information from a database that theuser did not know existed• Relationships between variables andcustomer behaviour that are non-intuitive isthe vital information that data mining extracts• Since the user does not know beforehandwhat the data mining process has discovered,it is a much bigger leap to take the output ofthe system and translate it into a solution for abusiness problem

Datamining tools provide answers to questions relatedto the detection of previously undetected patterns andare undirected in nature such as:

-Who are our best suppliers or most profitable customers? - Should we extend credit to a particular customer? - Which customers are likely to become profitable, when

and to what extent? - How do we optimally allocate resources to ensure

profitability and growth targets? - What are the root causes of quality issues and can we

cost- effectively minimize them? - What factors or combinations of factors are directly

impacting marketing campaigns?

• Intelligence is the aptitude to learn,comprehend, or to counter new or tryingsituations• It is the skillful use of reason and the capacityto apply knowledge to influence one'senvironment or to think conceptually• Business intelligence is a set of notions,methods, and practices, which improvesbusiness decisions. It uses information frommultiple sources and applies experience andassumptions that helps in understandingaccurately the intricacies of business dynamics.

• Business Intelligence (coined by

Gartner in the late 1980s) is “a user-

centered process that includes accessing

and exploring information, analyzing this

information, and developing insights and

understanding, which leads to improved

and informed decision making.”

• BI is the means by which organizations interpret thesea of organizational data to derive insights that arecritical to competing in the new economy• BI aids in:

- a deeper understanding of customer and partnerrelationships

- indicating key performance indicators- a consistent view of the organization from the executive

level to the front line• By translating these insights into action companiescan:

- increase profits- respond more quickly to changing market demands- improve accountability by giving every employee anaccurate view of the organization

The track - analyze - model decide –monitor loop is referred to as the closed loop model for business intelligence

• Track extracting, transforming, loading

(ETL), and integrating data into a data

warehouse as well as monitoring data in a

real-time or near real-time environment

• Transaction capturing systems or

operational systems capture data which is

later transformed, integrated, and loaded

into a data warehouse

• Analyze (analyzing data using BI tools)

- query and reporting, multi-dimensional analysis, and data mining

- Simple analysis methods like regression, co-relation , factor analysis etc.

are available in MS-EXCEL , ORACLE , SPSS, etc., .

- Data mining tools are available with software packages like SPSS, SAS,

Intelligent Miner, and Data Mind

• Model- formulating models for forecasting,optimization, and scenario planning- utilizing advanced analytics tools

• A model (a rule or a hypotheses) is madebased on the patterns discovered by datamining tools

•Decide

- arriving at a decision based on analysis and pre-

existing or newly developed models

- decision support systems use the models

developed as a result of data-mining and business

intelligence modeling processes for decision

making

•Act- a business manager uses the business analysis resultsto take an action (e.g., launching a new marketingcampaign based on the analysis of previous campaignresults, customer behavior, new promotional plan orinventory levels)- approving or denying a request for credit based onpast financial activity- re-negotiating sourcing contracts based on supplierdelivery trends, product quality, and warranty activitytrends, adjusting the type of data being tracked foranalysis, etc., .

• Identify buying behavior from customers

• Find associations among customerdemographic characteristics

• Predict responses to mailing campaigns

• Market basket analysis

• Detect patterns of fraudulent credit card use

• Identify loyal customers

• Predict customers likely to change their credit

card affiliation

• Determine credit cards spending by customer

groups

• Find hidden correlations between different

financial indicators

• Identify stock trading rules from historical data

• Claims analysis

• Predict which customers will buy newpolicies

• Identify behaviour patterns of riskycustomers

• Identify fraudulent behaviour

• Determine the distribution schedules

among

outlets

• Analyze loading patterns

• Successful BI architecture hasfour parts

- information architecture The informationarchitecture defines what business applicationsystems you need to access, report, and analyzeinformation to enable business decision making.

- data architecture The data architecture defines thedata, source systems and framework fortransforming data into useful information.

- technical architecture The technical architecturedefines the technology of the products andinfrastructure.

- product architecture The product architectureincludes the actual products used

Phase I Data Preparation:- Data Integration- Data Selection and Pre-analysis- Data Integration refers to the process of merging datawhichtypically resides in an operational environment having

multiplefiles or databases

Phase II Data Mining processor:- accesses a Data Warehouse that uses a relational databasesuchas DB2 for AIX/6000

- access is done through a standard SQL interface using amiddleware product which allows mining of data from

multiplesources

Phase III Presentation of facts and follow up:- presentation can be done in the Data Mining processor or

• a class of computer software built aroundmathematical models and algorithms(procedures) which, by converting data intoinformation and intelligence, help a managermake better decisions for his organization

• DSS are interactive computer based systems andsubsystems intended to help decision makersuse communication technologies , data ,documents , knowledge and/ or models tosuccessfully complete decision process tasks

• DSS can be divided into five basic tasks:- communications-driven DSS- data-driven DSS- knowledge-driven DSS- document-driven DSS- model-driven DSS