Article Preview
Top1. Introduction
Content on the web is billowing and growing at a faster rate. Search engines are productive tools to search the relevant information from the web. Due to the rapid growth of information over the internet, the task for finding the relevant information has become very difficult for every individual search engine. The success or failure of a search engine is unswervingly reliant upon the user’s satisfaction. The search engine users expect the information to be rendered to them in a small period of time. Users also expect that the results must be relevant and appropriate (Satya Sai & Raghavan, 2001). Most of the time the results returned by the search engines cannot entirely satisfy the requirement of the user and the search results are not very accurate and appropriate (Li et al., 2001).
Recommending relevant information is measured as an imminent factor in searching process nowadays, because current recommendation process is not based on user interest and is still following the predefined and static patterns of retrieved information in-spite of the fact that the resultant information needs to be filtered to meet user’s interest and objectives (Gulzar et al., 2019).The lack of any explicit arrangement and a wide variety of data available on the World Wide Web creates a challenge for its users to find the concerned data without extra efforts or without any outer help. It was believed that an individual general-purpose search engine lacks processing capability to cope-up with the amount of information being loaded on web nowadays (Sugiura & Etzioni, 2000; Manning et al., 2008).
There is a Meta-search engine concept which is gaining popularity among users and is built on top of other search engines. The user query in Meta-search engine is run across different components simultaneously, the result generated is ranked and best one is provided to the user (Meng, Yu & Liu, 2002). The primary aim of Meta-search engine is to overcome the inherited differences of individual search engines, and thus provide the finest result from the best search engines. Meta-search engine filters the top N results from individual search engine result and that’s why it is able to provide the most inclusive result set which is available on WWW. The traditional search engines crawl the web to retrieve the information, but on the other hand Meta-search do not crawl to provide the search result to the user. The Meta-search send the user query to dissimilar individual search engines at a time and only the top N filtered resultant documents are then visualized by the user in a window.
Meta-search engine too poses a few distinctive challenges in terms of the information which is not similar it gathers from individual engines. The outcome of the search consists of document ranking by individual engines which are also accompanied by a document title, a Snippet and a URL (Fabrizio, 2009; Rashid, 2008). But there are prominent advantages of Meta-search engines against individual engines by increasing the searching coverage on the web to provide high recall. It also increases the information retrieval effectiveness by increasing the precision and solves the problem of scalability of web-search. Search engines help to retrieve the relevant document links, but it is necessary to analyze the validity of the document, website and of the links. The existing search engines fail to apply the cognitive reasoning on the links and the semantics of the text. As a part of Meta-search engine, intelligent algorithms are required to select the relevant links. This research work applies the hierarchical clustering and fuzzy reasoner to club the links and rank the accurate results to cater the need of user requirement. This research work adapts the layered framework for retrieving the links from the web.