Web Objects Clustering Through Aggregation for Enhanced Search Results
Pushpa R. Suri, Harmunish Taneja · 2011
World Wide Web offer a rich mix of new challenges and opportunities to information computing researchers. The conventional search engine always returns a set of web pages in answer to a user query. Millions of web pages from organizations, institutions and personnel are made public electronically. With the web explosion and never ending raise of digital data, an added effect is the difficulty to retrieve relevant and reliable information from the Web. It is almost impossible for the naive user to get the right information in the answered search results as there is too much unrelated and out dated. The reason for this is rooted deep in the methodology for conventional Information computing on web that supports the indexing granularity for search as a web page. Search engines basically hunt for the potential web pages of user interests. On the contrary the user perspective in today's changing era is the information of a certain 'object' may be in the form of a cluster containing only the relevant data related to the object of interest rather than a tedious list of search results containing all the related and unrelated web pages. The similar theory can be applied to the queries from the point of view of developer. It requires grouping web objects into classes based on their attributes and links. This paper proposes algorithm for clustering web objects into different classes based on their links and identify relations dynamically. Results confirm the efficiency of the proposed approach as the user gets a cluster that contains only objects of interest from all the linked web pages.