Loading...

A novel algorithm for using GA in concept weighting for text mining

Zaefarian, R ; Sharif University of Technology | 2006

541 Viewed
  1. Type of Document: Article
  2. Publisher: 2006
  3. Abstract:
  4. The importance of good weighting methodology in information retrieval methods - the method that affects the most useful features of a document or query representative - is examined.. Weighting features is the thing that many information retrieval systems are regarding as being of minor importance as compared to find the feature and the experiments are confirming this. There are different methods for the term weighting such as TF*IDF and Information Gain Ratio which have been used in information retrieval systems, the paper provides a brief review of the related literature. This paper explores using GA for concept weighting which is a novel application to the field of text mining It proposes a new algorithm for using GA in term weighting for text summarization process and then by deploying it as an appropriate developed prototype, the outcomes are analyzed and some conclusions for Information Retrieval are considered. The empirical results lead us to indicate that the index of RECALL shows a significant increase over Tf-idf but no significant increase in terms of the index PRECISION providing a sound basis to carry out more tests on different texts needs to provide greater confidence in the proposed algorithm
  5. Keywords:
  6. Data mining ; Feature extraction ; Genetic algorithms ; Search engines ; Text processing ; Concept weighting ; Term distribution ; Term weighting ; Text mining ; Information retrieval systems
  7. Source: WSEAS Transactions on Computers ; Volume 5, Issue 12 , 2006 , Pages 2992-2999 ; 11092750 (ISSN)
  8. URL: http://www.worldses.org/journals/computers/old.htm