Recommendation-Based Meta-Search Engine for Suggesting Relevant Documents Links

Abstract

The information available online is mostly present in an unstructured form and search engines are indispensable tools especially in higher education organizations for obtaining information from the Internet. Various search engines were developed to help learners to retrieve the information but unfortunately, most of the information retrieved is not relevant. The main objective of this research is to provide relevant document links to the learners using a three-layered meta-search architecture. The first layer retrieves information links from the web based on the learner query, which is then fed to the second layer where filtering and clustering of document links are done based on semantics. The third layer, with the help of a reasoner, categorizes information into relevant and irrelevant information links in the repository. The experimental study was conducted on a training data set using web queries related to the domain of sports, entertainment, and academics. The results indicate that the proposed meta-search engine performs well as compared to another stand-alone search engine with better recall.