Download presentation
Presentation is loading. Please wait.
Published byMyrtle Taylor Modified over 6 years ago
1
A Web Mining Platform for Enhancing Knowledge Management on the Web KOK-LEONG ONG WEE-KEONG NG EE-PENG LIM Center for Advanced Information Systems, Nanyang Technological University 012ITI12 Song Mi-Kyoung
2
0. Contents Introduction Current Mining Scenario Existing Architecture
Whoweda, Wiccap Platform for Web Mining Data Mining Kernel, Components, Application Interface Application Example A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
3
1. Introduction MS MSDN KM & DataMining Recorded w/o knowledge analyst
Time invariant and global reach KM & DataMining A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
4
1. Introduction Web data Issues Loosening KDD Process
Web data is semi-structures at best Web data is distributed Web data is owned by everyone Web data has privacy concerns Web data is online and interactive Loosening KDD Process XML connecting components A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
5
2. Current Mining Scenario
PMML Interchange format for DM Supports DM seven models Independent on the application, system and architecture A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
6
2. Current Mining Scenario
JSR 073 The JAVA Data Mining application programming interface OLE DB Microsoft SQL with Data Mining CWM The OMG’s Common Warehouse Meta-data A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
7
3. Existing Architecture
Inter-API calls, data transfer b/w components by XML Web logs, multimedia data, transaction records A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
8
3. Existing Architecture
Whoweda Web warehouse : brings, preprocessing and storing data & information WICS : data manipulation module WIMS : data mining Module A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
9
3. Existing Architecture
Wiccap generate mapping rules extraction of information with mapping rule Whoewda + Wiccap Web protocol & interface in data mining kernel & platform can access all Web data in a structured and location independent manner A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
10
4. Platform for Web Mining
Data mining Kernel Provide a consistent set of interface to the Data Mining components Provide Security features to the incoming data : algorithm’s no direct access, sharing data Monitor the change on a Web site A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
11
4. Platform for Web Mining
Data mining Components Interpreters : bridge b/w application interfaces and the rest, translate, language framework Algorithms : data mining algorithms PMML : model exchange, component communication Visualization : visualization of PMML A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
12
4. Platform for Web Mining
Application Interface Web services API : application programming interface ASP interface : application service providers KQML/FIPA : standard for agent communication, software agent A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
13
5. Application Example User 1. User’s keyword
2. Filtering, removal of unrelated doc. 9. Click documents 10. track the documents 3. Negotiate services 4. Initiate protocol Translate the request 5. Retrieve documents 6. Data mining task 7. Translate to PMML 8. Return to search agent A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
14
6. Summary With Whoewda, Web data can be harvested into structured records that can be queried using Whoewda’s Web query language while Wiccap materializes information on the Web into XML documents that can be structurally proceed Standardization Web warehousing, Web technologies, software agent, database A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
15
7. Q & A Thank you. A Web Mining Platform for Enhancing Knowledge Management on the Web Spring Semester Data Mining ITI12 Song Mi-Kyoung
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.