Diversification of web search results using post-retrieval clustering

Diversification of results in web search engines is a very attractive area for researchers now a days. Information retrieval techniques mainly focus on the relevance of the documents retrieved but these techniques often fail to satisfy each user. In this work, we present a coverage based diversification using post retrieval clustering. We model clusters corresponding to the query based on the features of the web pages such as web pages of similar features are to be in one cluster and web pages from dissimilar features are to be in different clusters. A query can retrieve relevant and diverse result set if all the results cover as many clusters as possible.