New filtering framework for web personalization search / Anitawati Mohd Lokman and Aishah Ahmad

• Main Authors: Mohd Lokman, Anitawati, Ahmad, Aishah
• Format: Research Reports
• Language: English
• Published: Research Management Institute (RMI) 2012
• Online Access: http://ir.uitm.edu.my/id/eprint/17741/2/LP_ANITAWATI%20MOHD%20LOKMAN%20RMI%2012_5.pdf; http://ir.uitm.edu.my/id/eprint/17741/

This research attempts to design a filtering framework that incorporates the advantage of human based criteria, in addition to the computerized selection in order to produce more relevant and personalized search result to enhance the effectiveness of search result ranking technique. Currently, results suggested by search engine don't match user's expectations, which cause frustration to users. Previous research have shown only 20% to 45% of the common search results are relevant to the user's search. The search become harder when the keyword used is a homographic word. In search engine, page ranking can be affected by homographic keyword. For instance, the word 'Handicap' is one of the examples of homographic keyword. From initial observation with the use of Google search engine, there is two top search results for 'Handicap' keyword; i) disability, and ii) method of leveling any sports or games such as in golf. This shows that search engine will rank pages using general user's interest, and neglect the aspect of personalization involving user's interest. Thus, this research investigates common types of data to be used for personalization, designs framework based on types of data, and develops search cluster based on the data. The frame work is then validated with a demonstration of its implementation. This research will perform two phases of survey. The first survey is for data collection and the second survey is for validation of the filtering framework. This research presents new knowledge to enhance search result and contribute to the body of knowledge in Mathematics and Computing area. It enhances the search result by personalization that matches the user behavior, interest and ontology of metadata with other users. Also, by humanizing the search result it will lessen the gap between human and computer. The Filtering framework provides foundation in developing the algorithm for searching tools. This algorithm can be implemented in both website and search engine.