Enhancing the Search Results through Web Structure Mining Using Frequent Pattern Analysis and Linear Correlation Method

In this e-world, information is growing at an enormous rate due to voluminous amount of data is updated every day on the WWW. Moreover the popularity of web grows immensely as it covers wide spectrum in all domains ranging education, business, health, retail and surveillance. As the web serves as a connected medium to propagate information quickly with less cost, everyone depends on the WWW for retrieving relevant information for strategic decision making as well for their day to day operations. During these information retrieval, most of the users loss temper in navigating more number of links without getting the exact information catering to their needs. These issues create an interest among researchers to work on the area of Web mining. Generally, web mining is categorized into web content mining, web structure mining and web usage mining. This research work mainly focuses on web structure mining to retrieve the relevant links through frequent pattern analysis and linear correlation method. In this proposed work, based on the user query adjacency matrix is constructed for the extracted documents (links) and then parent document(link)) among these input documents is found out using frequent pattern analysis . Then statistical algorithm based on linear correlation is applied between the parent document and the remaining input document to rank the relevant links. The experimental results revealed that the performance of proposed approach in terms of precision and accuracy is high which improves the quality of search results.