GPODS

1
GPODS

• David Luper
• Delroy Cameron

2
Out Line

• General Project Overview
• GPS Intro
• System Architecture
• Data Mining and Semantic Analysis
• Example Queries
• Future Research
• Demo

3
General Project Overview

• Association Networks
• Our Plan of Attack
• Data Mining
• RDF Metadata
• What we hope to accomplish (analyzing GPS Data)

4
GPS Intro

• Gives every centimeter on earth a unique address
• Latitude Line
• Longitude Lines
• The Decimal Format System

5
GPS Intro
6
System Architecture
7
Database Schemas
8
Indexes

• TargetID, DateTimeStamp
• The primary key allowing for retrieval of a
  particular persons coordinates at a specified
  slice of time.
• Latitude, Longitude
• Allows for querying the entire database
  concerning who has ever been at a specific
  location.
• DateTimeStamp
• Allows for fast retrieval of where everyone in
  the database was at a specific time.
• DateTimeStamp, Latitude, Longitude
• Allows for retrieval of who was at a specified
  location at a specified time.
• ID, Latitude, Longitude
• Allows for querying all the times a person was at
  a specific location.

9
Data Mining and Semantic Analysis

• Binning (Discreet Map)
• Association Rule Mining

10
Data Mining and Semantic Analysis

• GPODS Ontology
• Query Processing
• Spatial
• Temporal
• Semantic Associations

11
GPODS Ontology Schema
gpodsgroup_name
foafPerson
gpodsGroup
gpodstarget_group
rdfsLiteral
subClass_Of
gpodsTarget
rdfsLiteral
rdfsLiteral
gpodsright
gpodsleft
gpodsRegion
gpodstarget_name
gpodstarget_group_status
rdfsLiteral
rdfsLiteral
gpodstop
gpodsregion_name
gpodstarget_gender
gpodsbottom
rdfsLiteral
rdfsLiteral
rdfsLiteral
rdfsLiteral
Figure1. GPODS Ontology Schema
12
Targets Ontology

• ltgpodsTarget rdfabout"LomezIniray"gt
• ltgpodstarget_namegtIniray Lomezlt/gpodstarget_
  namegt
• ltgpodstarget_gendergtFlt/gpodstarget_gendergt
• ltgpodstarget_group rdfresource"Terror_Grou
  p_Green"/gt
• ltgpodstarget_group_statusgtMemberlt/gpodstarge
  t_group_statusgt
• lt/gpodsTarget gt
• ltgpodsTarget rdfabout"PersonRandom198"gt
• ltgpodstarget_namegtRandom Person
  198lt/gpodstarget_namegt
• ltgpodstarget_gendergtFlt/gpodstarget_gendergt
• ltgpodstarget_group rdfresource"Good_Citize
  n"/gt
• ltgpodstarget_group_statusgtMemberlt/gpodstarge
  t_group_statusgt
• lt/gpodsTarget gt

13
Groups Ontology

• ltgpodsGroup rdfabout"Terror_Cell_Blue"gt
• ltgpodsgroup_namegtTerror Cell
  Bluelt/gpodsgroup_namegt
• lt/gpodsGroupgt
• ltgpodsGroup rdfabout"Terror_Cell_Red"gt
• ltgpodsgroup_namegtTerror Cell
  Redlt/gpodsgroup_namegt
• lt/gpodsGroupgt
• ltgpodsGroup rdfabout"Terror_Cell_Green"gt
• ltgpodsgroup_namegtTerror Cell Green
  lt/gpodsgroup_namegt
• lt/gpodsGroupgt

14
Query Processing
Target Group Region Time
John Smith Student UGA Arch 3/22/2007 1200pm
John Smith Student Terror Cell Green 3/22/2007 300pm
John Smith Student 226 Hardman 3/22/2007 800pm
Jane Doe Professor 226 Hardman 3/22/2007 800pm
Mark Adams Student Terror Cell Green 3/22/2007 300pm
has_student
visited_place
Mark Adams
John Smith
Jane Doe
Table1. Semantic Associations in Database
15
Spatial Association
gpodsHerndonTyler
gpodsWhatleyAmber
gpodsRegion277878
gpodsHerndonTyler
gpodsPersonRandom190
gpodsRegion22986
gpodsRegion1549
gpodsRegion356189
gpodsCrowleyTaylor
gpodsWhatleyAmber
Figure5. Spatial Semantic Associations
16
Ranking

• Context
• Popularity
• Association Length
• Rarity
• Trust
• Subsumption

17
Example Queries

• Simulate an event
• Populate Discreet Map
• Find places visited by more than 1 person
• Export to RDF
• Semantic queries for temporal and spatial
  association
• Ranking semantic findings
• Association rule mining for probability score
• Combining scores for an overall temporal and
  overall spatial association score

18
Future Research

• Path prediction and association trend recognition
• Approximate association rule mining and fuzzy
  logic
• Time sequence neural networks
• FOAF integration
• Sex offender / child protection queries
• Path learning (smart phone meets contextual
  mapping)
• Social networking (Helio)

19
Demo

• We will show you a demo know prepare accordingly