Enhanced information extraction on multiple search engines using hybrid algorithm

Abstract

Currently, more than seventy percent of the people are using the newlineInternet to find and search for their desired data such as cooking recipes, newlinebeauty tips, favorite movies, symptoms of diseases, job details, etc. through newlinethe Search Engines (SEs). All the SEs are retrieving their results as a set of newlinedocuments but the most relevant results are not retrieved by the SEs to satisfy newlinethe user requirement. It is still a challenging task for the researchers. In this newlinethesis, Multiple Search Engines (MSE) concept is applied with effective newlinealgorithms to retrieve most relevant results to the input query based on the newlinecombination of the K-Means* algorithm and Particle Swarm Optimization newline(PSO).In order to retrieve the relevancy of the resultant documents, KMeans* newlinealgorithm is applied, which enhances the concept of traditional newlineclustering technique named K-Means algorithm. Unlike the K-Means newlinealgorithm newly proposed K-Means* algorithm accepts the documents from newlinevarious resources in every cluster. The SEs such as Google, Yahoo and Bing newlineare selected as the resources for this work, as they are the most commonly newlineused SEs. The Google uses its own PageRank algorithm to provide the search newlineresults, the Bing uses the Index-serving technology named Sidebar to deliver newlinethe relevant search results and the Yahoo uses a Search Engine Provider to newlineshow the search results which are referred from some other companies. These newlinethree SEs results are fused by using Segmentation Fusion (SegFuse) method newlineand provide the single result list. The final single result set is fed as input to newlinethe K-Means* cluster algorithm, which classifies the same into different newlineclusters. newline newline