Main Article Content
Web search applications represent user information needs by submission of query to search engine. But still the entire query submitted to search engine doesn’t satisfy the user information needs, as users may want to obtain information on diverse aspects when they submit the same query. From this discovering the numeral of dissimilar user search goals for query and depicting each goal with several keywords automatically become complex. The inference of user search goals can be very valuable in improving search engine importance and user knowledge. To efficiently reflect user information needs to generate a pseudo-document to map the different user feedback sessions. Clustering pseudo-documents by K-means clustering is computationally difficult and semantic similarity between the pseudo terms is also important while clustering. To conquer this problem proposed a FCM clustering algorithm to group the pseudo documents and it also measures the semantic data alignment between the pseudo terms in the documents. The FCM algorithm divides pseudo document data in dissimilar size cluster by using fuzzy systems. FCM choosing cluster size and central point depend on fuzzy model. The FCM clustering algorithm assembles quickly to a local optimum or grouping of the pseudo documents in well-organized way. Semantic data alignment between the pseudo terms is used for comparing the similarity and diversity of pseudo terms. Finally, experimental results measures the clustering results with parameters like classified average precision (CAP), Voted AP (VAP), risk to avoid classifying search results and average precision (AP). It shows FCM based system improves the feedback session's outcome than the normal pseudo documents.