Web Search Result Clustering using IFCWR Algorithm

Abstract

In the modern age of Internet, usage of social media is growing rapidly on internet, organizing the data, interpreting and supervising User generated content (UGC) has become one of the major concerns. Updating new topics on internet is not a big task but searching topics on the web from a vast volume of UGC is one of the major challenges in the society. In this paper we deal with web search result clustering for improving the search result returned by the search engines. However there are several algorithms that already exist such as Lingo, K-means etc. In this paper basically we work on descriptive-centric algorithm for web search result clustering called IFCWR algorithm. Maximum numbers of clusters are randomly selected by using Forgy's strategy, and it iteratively merges clusters until most relevant results are obtained. Every merge operation executes Fuzzy C-means algorithm for web search result clustering. In Fuzzy C-means, clusters are merged based on cosine similarity and create a new solution (current solution) with this new configuration of centroids. In this paper we investigate the Fuzzy C-means algorithm, performing pre-processing of search query algorithm and try to giving the best solution.