Transcript Spatial Data Mining Architecture and Technologies PPT

Spatial Data
Mining
Architecture
and
Technologies
Team 12
Hari Kishan Bandaru
Sneha Anand Yeluguri
Parimi VSPVSK
Overview











Introduction to Spatial Data Mining
Related Work
Process of Spatial Data Mining
Process of Visual Space Data Mining
Common Data Mining Architecture
Spatial Data Mining Architecture
Visualization Data Model
Technologies
Advantages and Future Work
Conclusion
References
2/21
What is Spatial Data Mining?


Non–trivial search for interesting and unexpected spatial
patterns
Non-trivial Search:



Interesting:



Large (e.g. exponential) search space of plausible hypothesis
Ex. Asiatic cholera: causes: water, food, air , insects,…;
water delivery mechanisms: pumps, rivers, ponds, wells…
Useful in certain application domain
Ex. Shutting off identified water pumps => saved human life
Unexpected:


May provide a new understanding of world
Ex. Water pump – Cholera connection to the “germ” theory.
3/21
What is not a Spatial Data Mining?

Simple querying of Spatial data

Testing a hypothesis via a primary analysis

Uninteresting or obvious patterns in spatial data

Mining of non-spatial data
4/21
Motivation







To find new spatial patterns
To understand new geographic process for critical
questions
To analyze the fast growing spatial data
To explore large number of geographic hypothesis
To reduce plausible hypothesis
To discover relationships between spatial and non
spatial data
To build spatial knowledge-bases
5/21
Problem Statement




Rapid development in the technology of spatial data
storage, query, display and analysis.
Accessing spatial data through access methods often
need technology to spatial reasoning, geographic
computing and knowledge of space showing.
Spatial data mining technology is used to convert
spatial information of geographic coordinates into
useful knowledge and effective tools.
Visualization technology is used to generate graphics
from complex multi-dimensional data to display
objective things and their intrinsic connections.
6/21
Related Work

Transformation of map information mode to an
equilateral mode that consists of formalization ,
cognition and transmission.

Geo Visualization aim is to provide an information
exchange and feedback mechanisms for the
users.

Summarized visualization technology into three
points: Feature Identification, Feature Comparison
and Feature Interpretation.
7/21
Process of Spatial Data Mining







Inputting Spatial Datum
Feature Extraction and Feature Database Establishment
Data Warehouses Establishment
Data Extraction and forming Case Set
Create and Train Data Mining Model
Evaluating the Mined Out Model, discovering hidden
knowledge
Knowledge Application
8/21
Process of Visualization Spatial Data Mining

Filter : Extracting data of interest

Mapping : Creating geometric primitives

Draw : Translate geometric primitives into image

Feedback : Display the image
9/21
Common Data Mining Architecture




The efficiency of the data
mining should be improved
Historical method cant be
obtained effective
utilization
Interoperability between
different systems is bad.
For different application
object, pertinence is not
strong.
10/21
Spatial Data Mining Architecture
11/21
Visualization Spatial Data Mining Model
 Statistical
Mapping
Technique
 Color
 Thematic map
visualization
techniques
12/21
Four TierJ2EE Technology

Applet Call Servlet:

Servlet Call Java Bean:
13/21
Four TierJ2EE Technology

JavaBean Call JAFMAS components:

Return the results from JAFMAS:
14/21
XML and Space Data Warehouse
Technology


Read data from data warehouse and generate xml
document with unified expression form.
Change XML into DOM object model serves for upper
accessing.
15/21
Advantages




Use of thematic map visualization technology
helped users to explore spatial data
Increase the data processing speed significantly
Made abstract data much easier to understand
The proposed J2EE four tier architecture had
resolved the synergic work between the layers of
prototype system and between the components
in the layer.
16/21
Future Work

Many existing theoretical and technical issues
should be further explored and studied.



Ex: spatial data mining in structured modeling
Treatment of uncertain information
Need to explore similarity measure techniques of
mining model produced by statistics, fuzzy logic ,
rough set methods.
17/21
Conclusion



Combination of visualization technology and spatial
data mining has helped for analysis of spatial data
exploration.
New process and architecture were presented for
spatial datum data mining based on data ware
house.
The characteristics for spatial datum were analyzed
and difference between spatial data and traditional
relationship data were analyzed.
18/21
References



1. He Yueshun, Li Xiang; A Study of Spatial Data Mining
Technique Based on Web; In the preceedings of
International Conference on Engineering Profession,
General Topics for Engineers (Math, Science &
Engineering); Page 1-4; 2009
2. Xiao Qiang, Yan Wei, Zhang Hanfei; Application of
Visualization Technology in Spatial Data Mining; In the
preceedings of International Conference on Computing,
Control and Industrial Engineering; Page 153-157; 2010
3. He Yueshun, Xu Wei; A study of spatial data mining
architecture and technology; In the preceedings of 2nd
IEEE International Conference on Computing &
Processing (Hardware/Software); Page 163-166; 2009
21/21
Queries
19/20
20/20