Spatial Data Mining: Accomplishments and Research Needs

70 Pages·2005·1.86 MB·English

(cid:39) (cid:36)
Spatial Data Mining: Accomplishments and
Research Needs
Shashi Shekhar
Department of Computer Science and Engineering
University of Minnesota
Sea Surface Temperature (SST) in March, 1982
(cid:38) (cid:37)
(cid:39) (cid:36)
Why Data Mining?
(cid:63) Holy Grail – Informed Decision Making
(cid:63) Lots of Data are Being Collected
Business Applications:
•
Transactions: retail, bank ATM, air travel, etc
–
Web logs, e-commerce, GPS-track
–
Scientific Applications:
•
Remote sensing: e.g., NASA’s Earth Observing System
–
Sky survey
–
Microarrays generating gene expression data
–
(cid:63) Challenges:
• Volume (data) (cid:192) number of human analysts
Some automation needed
•
(cid:63) Data Mining may help!
Provide better and custmized insights for business
•
Help scientists for hypothesis generation
•
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
1
(cid:39) (cid:36)
Spatial Data
(cid:63) Location-based Services
Ex: MapQuest, Yahoo Maps, Google Maps, MapPoint
•
Figure 1: Google Local Search (http://maps.google.com)
(cid:63) In-car Navigation Device
Figure 2: Emerson In-Car Navigation System (In Coutesy of Amazon.com)
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
2
(cid:39) (cid:36)
Spatial Data Mining (SDM)
(cid:63) The process of discovering
interesting, useful, non-trivial patterns
•
patterns: non-specialist
–
exception to patterns: specialist
–
from large spatial datasets
•
(cid:63) Spatial patterns
Spatial outlier, discontinuities
•
bad traffic sensors on highways (DOT)
–
Location prediction models
•
model to identify habitat of endangered species
–
Spatial clusters
•
crime hot-spots (NIJ), cancer clusters (CDC)
–
Co-location patterns
•
predator-prey species, symbiosis
–
Dental health and fluoride
–
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
3
(cid:39) (cid:36)
Location As Attribute
(cid:63) Location as attribute in spatial data mining
(cid:63) What value is location as an explanatory variable?
most events are associated with space and time
•
surrogate variable
•
critical to data analyses for many application domains
•
physical science
–
social science
–
(cid:63) Location helps bring rich contexts
Physical: e.g., rainfall, temperature, and wind
•
Demographical: e.g., age group, gender, and income type
•
Problem-specific
•
(cid:63) Location helps bring relationships
e.g., distance to open water
•
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
4
(cid:39) (cid:36)
Example Spatial Pattern: Spatial Cluster
(cid:63) The 1854 Asiatic Cholera in London
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
5
(cid:39) (cid:36)
Example Spatial Pattern: Spatial Outliers
(cid:63) Spatial Outliers
Traffic Data in Twin Cities
•
Abnormal Sensor Detections
•
Spatial and Temporal Outliers
•
Average Traffic Volume(Time v.s. Station)
180
160
10
) 140
d
n
u
o20
B 120
h
ut
o
S 100
(30
D
n I
o 80
ati
St40
W 60
5
3
I
40
50
20
60
0
50 100 150 200 250
Time
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
6
(cid:39) (cid:36)
Example Spatial Pattern: Predictive Models
(cid:63) Location Prediction: Bird Habitat Prediction
Given training data
•
Predictive model building
•
Predict new data
•
Nest sites for 1995 Darr location
0
10
20
30
Marsh land
40
Nest sites
50
60
70
80
0 20 40 60 80 100 120 140 160
nz = 85
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
7
(cid:39) (cid:36)
Example Spatial Pattern: Co-locations
(backup)
(cid:63) Given:
A collection of different types of spatial events
•
(cid:63) Illustration
(cid:63) Find: Co-located subsets of event types
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
8
(cid:39) (cid:36)
What’s NOT Spatial Data Mining
(cid:63) Simple Querying of Spatial Data
Find neighbors of Canada given names and boundaries of
•
all countries
Find shortest path from Boston to Houston in a freeway
•
map
Search space is not large (not exponential)
•
(cid:63) Testing a hypothesis via a primary data analysis
Ex. Female chimpanzee territories are smaller than male
•
territories
Search space is not large !
•
SDM: secondary data analysis to generate multiple plau-
•
sible hypotheses
(cid:63) Uninteresting or obvious patterns in spatial data
Heavy rainfall in Minneapolis is correlated with heavy
•
rainfall in St. Paul, Given that the two cities are 10 miles
apart.
Common knowledge: Nearby places have similar rainfall
•
(cid:63) Mining of non-spatial data
Diaper sales and beer sales are correlated in evening
•
(cid:38) (cid:37)
Spatial Data Mining: Accomplishments and Research Needs
9