Ranking Function Optimizaton Based on OKAPI and K-Means

Abstract

This paper presents the Re-ranking algorithm in the Medical information retrieval based on K-Means and OKAPI. A program is implemented to discover multiple relevant aspects of each search, group the results, and re-rank them. The Re-ranking algorithm is to be created diverse results within each query and to output the best results in comparison to the Gold Standard. In the pre-processing side, K-Means clustering algorithm is active on Information Retrieval system. The results indicate that re-ranking algorithm and K-Means clustering made the some difference of original OKAPI ranking score. We propose an explanation for these results that is based on an analysis of the specifics of the clustering algorithms and the nature of document data.