Hyperlink Recommendation Based on Positive and Negative Association Rules

The new HRS method for hyperlink recommendation based on positive and confined negative association rules is presented in the paper. Discovered with the new PANAMA algorithm rules are merged and used in the form of recommendation functions, both to assess the existing hyperlinks and to suggest new ones. Positively and negatively verified and new hyperlinks are presented to the content manager and can considerably facilitate the maintenance of the web site structure and its adjustment to user behaviour. The experiments confirmed the usefulness of the Hyperlink Recommender System (HRS) and in particular, of negative recommendations based on confined negative association rules. Keywords Recommender System, Association Rules, Negative Association Rules, Hyperlink Assessment, Web Usage Mining, HRS.

• Type: Journal
• Year: 2008
• Where: NGC
• Authors: Przemyslaw Kazienko, Marcin Pilarczyk
• Keywords: Communications, Hyperlink, Negative Association Rules, NGC 2008, Recommender System