Recently two famous web page ranking algorithms are HITs and Page Rank. But Page Rank computed and refreshed off-line and not relevant to query term so not suitable for concept searching and finding topic-related communities instead HITs and SALSA are outperforms. In this paper, we discuss about to mine topic-related communities of web pages by HITs (Hyperlink-Induced Topic Search) and improve version of HITs, WHITs (weighted HITs) algorithm, which is based on hyperlink structure of web with double weighting of links matched with query term. The HITs and the WHITs algorithms are eigenvector based techniques for discovering “authoritative” web pages. Information Retrieval (IR) utilizes term based weighting method to discover relevant documents for a given query. Web IR utilities such as search engines tend to additionally process these relevant documents through link structure analysis and find rank score for each document within result set and present users for improving rank score rate of top ranked results. Existing link analysis algorithms are using principal eigenvector of resultant rank matrix for ranking. The multi topic or polymorphic query, the dominant topic discovers the major fraction within top ranked results and the sub-dominant topics are demoted. The improved version of HITs known as WHITs approach for link analysis serves for both ranking and grouping of pertinent links effectively.