Dynamic Clustering For Web Mining

Abstract

In today s era, World Wide Web is defined as huge collection of hyperlinked interconnected documents and Internet is commonly used to operate this service. Generally, a tool called search engine is used to retrieve information from WWW. A vast repository of web documents is maintained by Search Engine and return thousands of web pages in response to a user query but most of the results are irrelevant. Therefore, the real necessity is optimizing search results and listing them on the basis of their relevancy. newlineThe optimization of search results has been carried out by researcher earlier. It has been observed that existing optimization carried out relevancy of web pages individually based on their hyperlinks, popularity and user s click. But these approaches has lack in find relevant information, manage database, get precise information and high perceived latency. As a result, sometimes developed optimization also provide irrelevant results at top positions that are different from the user s desired relevancy. Thus, there is a requirement to rank the retrieved web pages based on content s relevancy and managing the web page repository to get precise information with low perceived latency. newlineA novel Dynamic Searching System (DSS) has been proposed to optimize the search results. This work focuses on the above mentioned issues and provides the solution for them by applying developed hybrid ranking algorithm called Content s Weight based Page Rank to rank the web pages based on their relevancy. The techniques of web mining such as web structure mining and web content mining are used to analyse the content, links and other important factors to rank the search results based on relevancy. By doing so, the user gets the precise information against entered query. Another major problem to managing the web page repository has been solved by integrating the concept of dynamic clustering. A technique of clustering says hierarchal clustering helps to organize huge amount of search results or web pages in clusters under its respect