To analyse the limitations of traditional recommendation algorithms, we propose a collaborative filtering recommendation algorithm based on XML fuzzy data. Firstly, the method grasped and modelled the fuzzy attribute features of items and established the membership matrix of fuzzy attribute features, then used XML similarity calculation to extract the similarity between project fuzzy attribute features and combined it with the traditional collaborative similarity to get the comprehensive similarity. Finally, recommendations were made through this comprehensive similar. Experimental results show that, compared with the traditional collaborative filtering recommendation algorithm, the establishment of membership matrix of fuzzy attribute features solves the problem that the description of goods or items is usually fuzzy, which makes the calculation of neighboring items more accurate and improves the accuracy of the recommendation system. At the same time, when the new item has no user scoring information, it can be recommended through the similar of the fuzzy attribute features among items, thus effectively solve the cold-start problem.
[1]
Xu,et al.
Improved Collaborative Filtering Recommendation Based on Classification and User Trust
,
2016
.
[2]
Lotfi A. Zadeh,et al.
Fuzzy Sets
,
1996,
Inf. Control..
[3]
Liu,et al.
Combining User-Based and Item-Based Models for Collaborative Filtering Using Stacked Regression
,
2014
.
[4]
An Zeng,et al.
Behavior patterns of online users and the effect on information filtering
,
2011,
ArXiv.
[5]
Yu Hong,et al.
Algorithm to Solve the Cold-Start Problem in New Item Recommendations
,
2015
.