Collaborative filtering is the most successful and widely used recommendation algorithm in E-commerce recommender systems currently. However, it faces severe challenge of cold-start problem. To solve the new item problem in cold-start, a cold-start recommendation method based on dynamic browsing tree model is proposed. Firstly, user browsing records are transformed to Dynamic Browsing Tree (DBT) based on product categories of E-commerce website. Secondly, a fresh degree decay operator based on access time is designed, then an item category similarity between leaves of DBT and new item is proposed. Finally, an Interest Matching Degree (IMD) measure is designed to compute the matching degree between new item and dynamic browsing trees of all users, thus those users who have higher IMD than designated threshold will be chosen as target audience for new item. The experimental results show that the proposed method can efficiently realize new item recommendation for collaborative filtering cold-start.
[1]
John Riedl,et al.
Analysis of recommendation algorithms for e-commerce
,
2000,
EC '00.
[2]
Bradley N. Miller,et al.
Using filtering agents to improve prediction quality in the GroupLens research collaborative filtering system
,
1998,
CSCW '98.
[3]
John Riedl,et al.
E-Commerce Recommendation Applications
,
2004,
Data Mining and Knowledge Discovery.
[4]
Nathaniel Good,et al.
Naïve filterbots for robust cold-start recommendations
,
2006,
KDD '06.
[5]
John Riedl,et al.
Combining Collaborative Filtering with Personal Agents for Better Recommendations
,
1999,
AAAI/IAAI.
[6]
John Riedl,et al.
Item-based collaborative filtering recommendation algorithms
,
2001,
WWW '01.