An efficient web page prediction based on access time-length and frequency

The ranking of web page for the Web search-engine is one of the significant problems at present. This leads to the important attention to the research community. The vast crawling and indexing infrastructure is required for sustaining the accurate and timely web page ranking. In contrast, it is not possible to crawl the web continuously because of it non static nature, huge number of available web pages and bandwidth constraints. The same will be resulted for indexing and page ranking technique because its complex computation nature and high expense. Thus the requirement for new rank prediction technique arises. To provide better rank prediction, two biasing features are considered that can favors more significant web pages to the user. One feature is the length of time spent on visiting a page and the other feature is the frequency of the visited page. This technique uses first order Markov Chain scenario. Even though this technique possesses several advantages, it is memoryless to determine the weights of edges in the directed graph of a website and to expand the subgraph for a user´s current navigational path. In order to overcome this disadvantage, the higher order Markov Chain is implemented in this paper. The experimental result shows that the proposed technique for ranking the web page yields better prediction accuracy when compared to the existing techniques.