Developing methodology to improve the accuracy of popularity based ranking algorithms via implicit links

There exist high number of accessible documents on the Web and the number is growing exponentially. More and more people are turning to the Internet when searching for a particular information. However, the Web is tremendous resource and it must be used carefully and critically. Web Search engines are the primary tools for searching information. Web Search engines are faced with a number of difficult problems in enhancing the quality of their performance. It is a higly important problem to select best documents among the ones that match the query. We proposed a methodology to improve the accuracy of PageRank algorithm which is an example of popularity based ranking algorithms. In this methodology, we used the implicit links that are available in Web pages. In many Web pages, other Web pages are only mentioned by name without hyperlinks. These names are accepted as implicit links which are linked to other Web pages. The implicit links can affect the calculation of Web pages' popularity as direct links. Based on this, we incorporate the implicit links into Web graph consisting of hyperlinks and create an enriched Web graph. In our methodology, we improve the PageRank algorithms which is an example of polularity based ranking algorithms by considering the implicit links. Hence, we perform the calculations of improved PageRank scores on the enriched Web graph and investigate whether the inclusion of implicit links provide better ranking results than similarity based algorithms and plain PageRank algorithm. After user study, we show that improved PageRank algorithm gives better results in terms of precision and recall, compared with two other baseline algorithms. The results are promising and indicate that inclusion of implicit links improves the accuracy of pagerank computations.