Design and Evaluation of Semantic Guided Search Engine

Abstract

Search engines provide a gateway through which people can find relevant information in large collections of heterogeneous data. Search engines efficiently service the information needs of people that require access to the data therein. Web search engines service millions of queries per day, and search collections that contain billions of documents. As the growth in the number of documents that are available in such collections continues, the task of finding documents that are relevant to user queries becomes increasingly costly. In this work, Semantic Guided Internet Search Engine is built to present an efficient search engine – crawl, index and rank the web pages – by applying two approaches. The first one, implementing Semantic principles through the searching stage, which depends on morphology concept – applying stemming concept – and synonyms dictionary, and the second, implementing guided concept during input the query stage which assist the user to find the suitable and corrected words. The advantage of guided concept is to reduce the probability of inserting wrong words in query. The concluded issues in this research that the returned web pages are semantic pages yielding with synonyms depending on the query terms which achieved the concept of semantic search and as compared with Google, good results are appeared depending on the Recall and Precision measurements reaching 95% - 100% for some queries in spite of the differential of environment between the two systems. Also, the performance of the search is improved by using guided search and by using the improved PageRank, which reduces the retrieved time. Finally removing stop words from a document minimizes the storage space, which enhanced the proposed system.