Design to post-processing of IR based on fuzzy clustering analysis

Man has been exploring continuously to improve the precision and intelligent degree of IR in order to deal with large quantities of information. In traditional methods of IR, retrieval results are disposed so little that the quality of IR is not high. To overcome this issue, a clustering algorithm based on cut-set is designed in this paper. This algorithm carries through clustering based on no leading text choice to many retrieval results so that users only consider the relevant sub-classes and abandons the non-relevant data. Hence, users can browse fewer documents. Experiments on live data of text snapshot show this algorithm can improve the query efficiency of IR system considerably at the cost of much less time.