A MICROBLOG SUBSPACE ENSEMBLE CLUSTER APPROACH FOR PREDICTING USER INTEREST IN WEB MINING

Web search engines are designed to meet the demands of all users, regardless of their unique requirements. The project's goal is to create a customised online search engine that takes into account the user's interests and creates search results from web links based on the user's semantic profile. To organise web documents and add structure to the findings presented to the customer, the proposed scheme employs clustering and re-ranking algorithms. With a dynamic rank prediction method, this paper proposed an automated microblog subspace ensemble clustering algorithm (MSECA) based on entropy for discovering the interest trend over users' web logs (DRP). On the basis of the clustering algorithm, we implemented the knowledge entropy. Unlike conventional clustering algorithms, our approach does not enable the end user to specify any parameters. In the meanwhile, it will find clusters of any form or scale. The effectiveness of our algorithm in the issue of high-dimensional and non-informative prior pattern relation estimation has been demonstrated by experimental findings on real-world datasets. The specification also includes keywords that aid in the integration of the user's current interests. Finally, the experimental results reveal that the suggested search engine outperforms commercial search engines in various metrics.