Spatial Data mining - Spatial Database Group

advertisement
DISCOVERING SPATIAL CO-LOCATION
PATTERNS
PRESENTED BY: REYHANEH JEDDI & SHICHAO YU (GROUP 21)
CSCI 5707, PRINCIPLES OF DATABASE SYSTEMS, FALL 2013
11/26/2013
RELATION WITH THE COURSE IS CHAPTER 28 (DATA MINING )
Overview
Introduction
spatial data mining
Association Rule
Co-location Miner Algorithm
• Data mining is finding some methods in large data
sets and using stored data from data warehouse to
analyze and manage the data to reduce future
problems.
• Spatial Data mining is using the Data mining
methods for spatial data and reaches some designs in
data according to Geography  location, area and any
same aspect.
 Spatial data mining methods :
 spatial OLAP and spatial data warehousing : Multi
dimensional spatial databases
 Characterization of spatial objects : Compare data
distinctive
 Spatial organization: Rules for city
 Spatial allocation and indicator : Arrange countries
 Spatial clustering : Bundling homes
 Similarity analysis in spatial databases : Similar area
Spatial data role
 Analyzing level  connection and narrowing
 Location role  space’s phenomena
 Spatial databases
 large scale and datasets
 Spread domain : Ecology, Society safety , Health issues, ….
 Map’s images  Various time : 20 to 100
 Ecology  Co_accident
 Spatial design  Co_location pattern
•
Ecosystem data sets' spatial pattern :
•
Local co_location pattern
•
spatial co_location pattern
ASSOCIATION RULE

Association Rule--- analyzing and predicting
– An implication expression of the form X  Y, where X
and Y are itemsets
– Example:
{Milk, Diaper}  {Beer}

Rule Evaluation Metrics
– Support (s)

Fraction of transactions that contain both X and Y
– Confidence (c)


Measures how often items in Y
appear in transactions that
contain X
Given a set of transactions T, the goal of association
rule mining is to find all rules having
– support ≥ minsup threshold
– confidence ≥ minconf threshold
TID
Items
1
Bread, Milk
2
Bread, Diape r, Beer, Eggs
3
Milk, Diape r, Beer, Coke
4
Bread, Milk, Diape r, Beer
5
Bread, Milk, Diape r, Coke
Example:
{Milk , Diaper }  {Beer}
 (Milk, Diaper, Beer )
2
 0.4
|T|
5
 (Milk, Diaper, Beer ) 2
c
  0.67
 (Milk, Diaper )
3
s

Limitations of Transactions on Spatial Data
-Order sensitive transactions
- Transaction over space
-Support and confidence are ill-defined - a priori algorithm
-May under-count support for a pattern
-May over-counter support
Overview
Introduction
spatial data mining
Association Rule
Co-location Miner Algorithm
From Transactions to Neighborhoods
 Transactions
-discrete, Independent, disjoint
 Neighborhoods
-Continuous, Spatial related
An Event centric co-location model
table instance
3/4 2/5
2/5
2/4 2/3
2/4
3/5 2/3
3/5
Illustration: Co-location Miner algorithm
 Generate candidate co-locations
 Participation indexes calculation
 Co-location rule generation
Advantage to Other Mining Methods
• Event centric co-location model
– Robust in face of overlapping neighborhoods
• Co-location Miner algorithm
– Computational efficiency
– High confidence low prevalence co-location patterns
– Validity of inferences
REFERENCES
Book:
•
Introduction to Data Mining, By Pang-Ning Tan; Michael Steinbach; Vipin Kumar 6th Edition
Articles :
• http://edugi.uji.es/Bacao/Geospatial%20Data%20Mining.pdf
• http://www.spatial.cs.umn.edu/paper_ps/sstd01.pdf
• http://en.wikipedia.org/wiki/Data_mining
• http://www.docstoc.com/docs/121010850/Spatial-Data-Mining---PowerPoint
• http://www.spatial.cs.umn.edu/paper_ps/co-location.pdf
Pictures:
• http://www.spatial-accuracy.org/FromICCSA2008
• http://gcn.com/articles/2008/11/14/the-state-of-spatial-data.aspx
• http://www.ec-gis.org/Workshops/7ec-gis/papers/html/gitis/gitis.htm
• http://www.spatialdatamining.org/software
• http://www.spatialdatamining.org/
• http://www.geocomputation.org/2000/GC059/Gc059.htm
THANK YOU
Download