Trajectory Pattern Mining

Fosca Giannotti, Mirco Nanni,Dino Pedreschi, Fabio PinelliKDD Lab (ISTI-CNR & Univ. Pisa)

Presented by: Qiming Zou

Overview

Motivations Trajectory T-Pattern Regions of Interest Future Work Q&A

Motivations

Large number of mobile devices, mobile services available

Motivations It is possible to collect position traces fro

m such devices We can extract information and patterns

from these data to describe mobility behaviors

Use this information for fields such as urban planning

Trajectory

Trajectories are sequences that contain the spatial and temporal information about movements

Trajectory Trajectories are usually given as

spatiotemporal (ST) sequences: <(x0, y0, t0), ..., (xn, yn, tn)>

xi, yi is the position coordinate relative to the origin

ti is the time stamp for the position information

Trajectory 2D and 3D representation of a

trajectory:

T-Pattern A Trajectory Pattern (T-Pattern) is a coup

le (s, α), where: s = <(x0, y0),..., (xn, yn)> is a sequence of n+1 lo

cations α= <α1,..., αn> are the transition times such th

at αi = Δti = ti – ti-1

T-Pattern A T-Pattern Tp occurs in a trajectory if it c

ontains a subsequence S such that: each (xi, yi) in Tp matches a point (xi’, yi’) in

S the transition times in Tp are similar to those

in S

T-Pattern The same exact spatial location (x, y)

usually never occurs Yet, close locations often represent the

same place The same exact transition times

usually do not occur often However, close times often indicate

similar behavior

T-Pattern To solve the problem, we introduce

the notions of: Spatial neighborhood: Two points match

if one falls within a spatial neighborhood N() of the other

Temporal tolerance: Two transition times match if their temporal difference is ≤ τ

T-Pattern Example:

Regions of Interest It is too computational intensive and

yield little practical use to generate all T-Patterns

Solution: Use a Regions of Interest approach, only use these regions as nodes of the T-Patterns

Regions of Interest Given a set of Regions of Interest R, defin

e the neighborhood of (x, y) as:

Neighbors = belong to the same region Points in no region have no neighbors

Regions of InterestS=<(x0, y1, t1), ..., (x4, y4, t4)>

=>

<(R4, t0), (R3, t2), (R3, t3), (R1, t4)>

Regions of Interest What if the Regions of Interests are

not known before hand? Define heuristics for automatic

Regions of Interest extraction from data: Geography-based (crossroads) Usage-based (popular places) Mixed (popular squares)

Future Work

Application-oriented tests on large, real datasets

Study relations with Geographic background knowledge Privacy issues Reasoning on trajectories and patterns

Trajectory Pattern Mining

Questions?