Jaccard Snowball Robust Regression Based Page Ranking With Big Data In Dynamic Web Environments

Abstract

World Wide Web (WWW) comprises a large volume of information and provides access to the users at anyplace and anytime. By increase of resources, users have several difficulties for detecting valuable data. But, web information methods are focused on relevant data but it still not concentrated on pages interesting to the users. In order to find the interesting web pages of the user, a novel machine learning technique called Jaccard Snowball Preprocessed Tanimoto Robust Regressive Page Ranking (JSPTRRPR) method is introduced. JSPTRRPR method was used for enhancing accuracy of web page ranking with minimum time consumption. In order to achieve this contribution, the JSPTRRPR method includes two major processes namely preprocessing and page ranking.  WWW comprises a number of information. The JSPTRRPR method considers the number of user query as an input. Jaccard Snowball Stem Preprocessing Model is applied for eliminating stop as well as stem words from the input user query. After that, Tanimoto Robust Regressive Page Ranking Model is carried out to rank the significant pages at the top based on the user queries. Robust regression analysis used to find the top-ranked web pages by measuring theTanimoto similarity between the user query and webpage contents. The web pages with higher similarity value get ranked at the top using the Borda count fractional ranking method. Finally, a series of experimental results has signified and confirmed that the JSPTRRPR method achieves prominent performance in terms of higher ranking accuracy and minimum false positive rate as well as ranking time by number of user query.