基于标签和因子分析的协同推荐方法

根据在线社区中群体的历史行为进行物品(或信息)推荐是当前研究热点之一,传统推荐算法都面临数据稀疏性问题的挑战. 针对传统推荐算法知识表示的局限性进行了研究,提出了一种基于标签系统的用户行为知识表示法,把用户在物品上历史行为的统计,转化为对用户在物品标签上的统计,从而缓解数据稀疏的情况. 为了降低标签维度过高导致的计算复杂性问题,提出了采用因子分析法,抽取出潜在重要且稳定的特征因子向量来最终表示用户的历史行为,并据此度量用户行为在特征因子向量上的相似性. 最后采用协同过滤的思想给出了一种新的协同推荐方法. 通过在真实数据集上的大量对比实验,表明该方法在处理具有稀疏性的数据集时,总是能保持更高且更稳定的推荐准确率.

Collaborative Recommendation Method Based on Tags and Factor Analysis

Item (or information) recommendation is one of hot research topics currently. However the is-sue of sparseness in dataset challenges all traditional recommendation algorithms. Limitations of knowl-edge representation in traditional recommendation algorithms were studied. The tag-system-based knowl-edge to represent information of each user’s behavior was proposed. That it the account on user’s behav-ior on items is transferred to an account on a user’s behavior on tags. To decrease the computation com-plexity on high dimensional tag-based datasets, a factor analysis method was taken to extract those most important latent factors to represent users’ behaviors. Based on each user’s representing vector of latent factors, a new way was given to compute similarities among users. By incorporating this similarity meas-ure, a new collaborative recommendation method with low sensitivity to sparseness was built to meet the need of practical and dynamic datasets. Experiments were carried on real-world datasets to compare the proposed method with other state-of-the-art collaborative filtering and matrix factorization based recom-mendation methods. It is shown the proposed method can achieve better prediction accuracy while keeps a lower sensitivity to sparseness.