Development of hybrid similarity measure using fuzzy logic for performance improvement of information retrieval system

The performance of information retrieval is dependent upon how effectively the documents can be ranked according to numeric similarity measure between the query and the document. The cosine and Jaccard are commonly used similarity measures. The authors have presented new similarity measure by combining Cosine and Jaccard similarity measures using fuzzy logic. The experiments are performed on CACM data collection. The proposed hybrid similarity measure gives better results than other two similarity measures.