PRESENTED BY:
T.Youva Vyshnavi.

---

[attachment=11822]
Web Mining
Web Mining is the application of data mining techniques to discover patterns from the web.
Introduction
• There are many web search engines exit.
• Which are irrelevant to browser.
• Personalized web search takes key words from user as expression set .
• And deliver results from personalized indexing server of their community/ organization to determine the relevance of pages.
Objective
Improves effectiveness of information retrieval
Approach proposed is community based personalized web search
Depends on personalized search strategies.
In this process
– communication usage of the users is tracked
– index the results and
– deliver there indexed results with the top ranks in the further searches
Existing System and its limitations
The personalized search based on user profiles.
It has various limitations.
• Information provided by a user himself/herself is only used to create user profiles.
• Algorithms work only for repeated queries.
• Suggested for search engines where security is not needed.
• personalization not work properly under all situations.
• Queries with low click entropy value is worse than generic search.
• In terms of Short term interest-based search personalization is no need.
Not considering re-ranking of results based on community/organization/domain
Proposed System
• Automatically predict query which will benefit from previous personalization algorithm.
• Involves communities/organization/domain.
• Improves effectiveness.
 Precision
 Recall
• Gives better performance than simple personalization algorithm.
• Tracks user usage communication.
• Index results.
• Re-rank results based on communities.
Flow
• Planning to implement the community based personalized search algorithm.
• User query is taken as input.
• Input given to search model- Search crawler- Search engine
 Crawl web for results.
 Accept results prone them based on semantic relevance.
 Deliver the results.
 Track the communication usage.
 Index the results most frequently used.
 Deliver there indexed results with top rank in further search
Conclusion
• Planning to implement the community based personalized search algorithm.
• Performs better than simple personalization technique.
• Gives better performance than user profile.
• No need of providing any explicit user details by him/her selves as feed back.
• Index results & re-rank them based on community.
• Results in improved retrieval performance.
• Which solve the problem of retrieving generic results to user.