spatial data mining- applications

25
Spatial Data Mining- Applications Hemant Kumar Jerath,B.Tech. MS Project Student Mangalore University Advisors: Dr. B.K Mohan & Dr.(Mrs.).P. Venkatachalam CSRE, IIT Bombay

Upload: madison

Post on 10-Jan-2016

89 views

Category:

Documents


6 download

DESCRIPTION

Spatial Data Mining- Applications. Hemant Kumar Jerath, B.Tech. MS Project Student Mangalore University Advisors: Dr. B.K Mohan & Dr.(Mrs.).P. Venkatachalam CSRE, IIT Bombay. Beyond Conventional GIS Analysis: Spatial Analysis, Geospatial Data Mining, and Knowledge Discovery. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Spatial Data Mining- Applications

Spatial Data Mining-Applications

Hemant Kumar Jerath,B.Tech.

MS Project Student

Mangalore University

Advisors: Dr. B.K Mohan & Dr.(Mrs.).P. Venkatachalam

CSRE, IIT Bombay

Page 2: Spatial Data Mining- Applications

Beyond Conventional GIS Analysis: Spatial Analysis, Geospatial Data Mining, and

Knowledge Discovery

Page 3: Spatial Data Mining- Applications

Conventional GIS Operations

Retrieve Spatial and Attribute Data

Measure

Descriptive Statistics

Classification

Overlay

Buffer

Network Analyses

Page 4: Spatial Data Mining- Applications

Where do minorities live in New Jersey?

Which census tracts have percent minority > 25%

Conventional GIS Operations: Attribute Query

Page 5: Spatial Data Mining- Applications

Conventional GIS Operations: Attribute Query

SQL

Structured (Standard) Query Language

Formal language for interacting with relational databases

SELECT <fields>

FROM <tables>

WHERE <condition>

SELECT %Min

FROM CensusTracts

WHERE %Min > 0.25

Page 6: Spatial Data Mining- Applications

Tougher GIS Operations: Finding Spatial Relationships in Very Large, Spatio-Temporal, Multi-Dimensional Data Sets

Page 7: Spatial Data Mining- Applications

Some Tougher Questions…

1. Are hazardous facilities clustered in certain areas? If so, what are the socioeconomic characteristics associated with these clusters?

2. What is the influence of degree of urbanization, proximity to transportation, and industrial land use on the relationship between locations of hazardous facilities and the racial and other socioeconomic characteristics of the communities that host these facilities?

Page 8: Spatial Data Mining- Applications

Exploratory Spatial Data Analysis (ESDA)

The sophisticated quantitative analysis of spatial and spatio-temporal data, typically involving interactive and dynamic interfaces.

Related terms:

•Spatial Statistics

•Spatial Analysis

Page 9: Spatial Data Mining- Applications

A process to discover hidden facts and useful information contained in databases.

Related terms:

•Machine Learning

•Pattern Recognition

Geospatial Data Mining (DM) and Knowledge Discovery (KD)

Page 10: Spatial Data Mining- Applications

Some ESDA, DM, and KD Operations

Finding Spatial and Multi-Dimensional Clusters

Summarizing Variables by Other Variables

Finding Associations Among Attributes

Predicting Values

Feature Extraction

Page 11: Spatial Data Mining- Applications

GeneralizationDr.Hans,SMU

• 500 weather probes– Monthly mean temp and precipitation

• Weather pattern of season 1990

• 18,000 tuples– Attribute Induction

• Non-spatial(merging of tuples based on non-spatial concept hierarchies)

• Spatial(merging of spatial objects based on the concept hierarchies-spatial region merging and/or spatial approximation)

Page 12: Spatial Data Mining- Applications

Spatial Associative Rules

• Finding large towns and nearby objects like mines, country boundary, water(sea, lake, or river) and major highways.– Generalisation of the above spatial objects

Page 13: Spatial Data Mining- Applications

Spatial Clustering

• A request to all expensive houses in an area.

• Finding down the relationships of the clusters with other spatial objects like roads, different land-use area

Page 14: Spatial Data Mining- Applications

Clustering<house type,price,size>

Distribution

Page 15: Spatial Data Mining- Applications

Spatial Data Mining: A Database ApproachMartin Ester, Hans-Peter Kriegel, Jorg Sander

• Step I: Discover centers based on some non-spatial attribute[clustering-descriptive mining]

• Step II: determine the (theoretical) trend of some non-spatial attribute.

• Step III: discover the deviation of the theoretical trends

• Step IV: explain the deviation by the spatial object, e.g. may be presence of some infrastructure.

Page 16: Spatial Data Mining- Applications
Page 17: Spatial Data Mining- Applications
Page 18: Spatial Data Mining- Applications

Spatial Classification

• Non-spatial attribute e.g. no. of salespersons in a store

• Spatially related attribute with non-spatial values, like population living within 1km from store

• Spatial predicates, like – Distance_less_than_10km(X,a)

• Spatial function, like driving_distance(X,beach)

Page 19: Spatial Data Mining- Applications

Decision Tree

Description of classified objects

Description of census block group

Buffers are definedFor Trade Area

Page 20: Spatial Data Mining- Applications

Classification is developed using ID3 algorithm

Page 21: Spatial Data Mining- Applications

High_profit=N

High_profit=Y

Page 22: Spatial Data Mining- Applications

Geo-spatial Data Mining and KDD using Decision Tree Algorithm-A case study of soil data sets

Jianting Zang, Diansheng Guo, Qing Wan

Page 23: Spatial Data Mining- Applications

Association Rule Mining on RSI using P-TreesQin Ding, Qiang Ding, William Perrizo

• Identification of high and low crop yield

• Insect and weed infestations

• Nutrient requirement

• Flood damage assess

Page 24: Spatial Data Mining- Applications

ILP, SPADA(Spatial Pattern Discovery Algorithm)Donalto Malerba, Francesca A Lisi

• Find associations between– Reference objects (Towns)

– Task Relevant Objects ( road network, hydrography and administration layers)

• Stockport Census data– Socio-Economic phenomenon

– Census data (80 tables, 120 attributes)

Page 25: Spatial Data Mining- Applications