Abstract: “In this paper, we propose a method of Multi- Objective Genetic Algorithms (MOGAs), NSGA-II and SPEA2, for document clustering with semantic similarity measures based on WordNet. The MOGAs showed a high performance compared to other clustering algorithms. The main problem in the application of MOGAs for document clustering in the Vector Space Model (VSM) is that it ignores relationships between important terms or words. The hierarchical structure of WordNet as thesaurus-based ontology is an effective technique, which is used in semantic similarity measure. We tested these algorithms on Reuter-21578 collection data sets and compared them with Genetic Algorithms (GA) in conjunction with the semantic similarity measures based on WordNet. Also, we used F-measure to evaluate the performance of these clustering algorithms. The experimental results show that the performance of MOGAs based on WordNet is superior to those of the other clustering algorithms in the same similarity environments.”

Keywords: Document Clustering, Multi-Objective Genetic Algorithm, Semantic Similarity Measure, WordNet