Kabir, Shaily (2012) New Similarity Measures for Capturing Browsing Interests of Users into Web Usage Profiles. Masters thesis, Concordia University.
Kabir_MCompSc_F2012.pdf - Accepted Version
Available under License Spectrum Terms of Access.
The essence of web personalization is the adaptability of a website to the needs and interests of individual users. The recognition of user preferences and interests can be based on the knowledge gained from previous interactions of users
with the site. Typically, a set of usage profiles is mined from web log data (records of website usage), where each profile models common browsing interests of a group of like-minded users. These profiles are later utilized to provide
personalized recommendations. Clearly, the quality of usage profiles is critical to the performance of a personalization system. When using clustering for web mining, successful clustering of users is a major factor in deriving effective usage profiles. Clustering depends on the discriminatory capabilities of the similarity measure used. In this thesis, we first present a new weighted session similarity measure to capture the browsing interests of users into web usage profiles. We base our similarity measure on the reasonable assumption that when users spend longer times on pages or revisit pages in the same session, then very likely, such pages are of greater interest to the user. The proposed similarity measure combines structural similarity with session-wise page significance. The latter, representing the degree of user interest, is computed using page-access frequency and page-access duration. Web usage profiles are generated by applying a fuzzy clustering algorithm using this measure. For evaluating the effectiveness of the
proposed measure, we adapt two model-based collaborative filtering algorithms for recommending pages. Experimental results show considerable improvement in overall performance of recommender systems as compared to other known similarity measures. Lastly, we propose a modification by replacing structural similarity by concept (content) similarity, which we expect would further enhance recommendation system performance.
|Divisions:||Concordia University > Faculty of Engineering and Computer Science > Computer Science and Software Engineering|
|Item Type:||Thesis (Masters)|
|Degree Name:||M. Comp. Sc.|
|Date:||6 June 2012|
|Thesis Supervisor(s):||Mudur, Sudhir P. and Shiri, Nematollaah|
|Keywords:||web personalization, web usage mining, recommender system|
|Deposited By:||SHAILY KABIR|
|Deposited On:||30 Oct 2012 15:46|
|Last Modified:||05 Nov 2016 02:11|
Repository Staff Only: item control page