Design a Perception Based Semantics Model for Knowledge Extraction

Abstract

Perception based semantic search systems are common in modern era but accommodating location based searching is rare in such search engines. These works consider semantic perception based web searching with location sensitive mechanism. Web mining and interpretation is classical problems of web. The proposed thesis works towards this aspect with incorporation of word sensing. The motive of study is to give reliable and specific result to the user. Entire work of thesis formation is partitioned into phases. One of the crucial phases of study is data collection. Data collection phase uses JOC tool and with the help of this 500 websites are stored within local database. Second phase is getting URL from the user and storing the location of user. This is accomplished using Google-API service. Third phase is oriented towards pre-processing user query and eliminating stop words from the user. Forth phase uses word dissemination mechanism to check for the relevant alternatives present within the dictionary. The unique mechanism of user login is used in order to provide location sensitive information to the authenticated users. In addition, local buffer is maintained in order to provide quick search for relevant content using hash map function. Recommendations are handled using content and collaborative filtering mechanism. Result comparison is presented in terms of execution time, number of URLs fetched and patterns extracted. Result is improved by the margin of 10% that is significant enough to prove worth of study. newline