融合社会网络和项目特征的移动应用推荐

App Recommendation Combing Item Features and Social Networks

随着移动应用的数量增长,如何在海量应用中为用户推荐其感兴趣的应用受到了广泛关注.传统的协同过滤算法通过提取用户共同评分项目信息来计算用户相似度.然而,协同过滤算法普遍存在数据稀疏性问题,这在一定程度上导致了Pearson公式的计算结果不能准确的反映用户的相似程度.为了改善由于数据稀疏性问题导致的推荐结果不准确,我们使用K-means方法对项目进行基于语义相似的聚类,以实现基于相似项目的用户相似度计算,在此基础上,提出一种融合社会网络和项目特征的移动应用推荐.实验表明:融合了社会网络和项目特征的移动应用推荐能够缓解数据稀疏性问题对协同过滤算法的不利影响,在一定程度上提高了推荐结果的准确度.

;In recent years, the number of mobile applications has increased as the number of bamboo shoots after a spring rain. With the increasing number of mobile applications ,how to find the users＇ favorite mobile applications has been a more and more important issue. Both domestic and foreign scholars are focused on how to improve the accuracy of the recommender system. Collaborative filte- ring algorithm is widely used in all kinds of recommender systems. The traditional collaborative filtering algorithm calculates the user＇s degree of similarity by extracting the information of common score project. However, the sparse of user-item ratings is a common problem in collaborative filtering recommender systems. That is to say in the score matrix,the ratio of common rated items is less. So the accuracy of recommendation will be influenced because users with similar preferences can＇t be found accurately. Many research pa- pers are focused on how to increase the accuracy of recommender systems by alleviating the sparse of user-item ratings. In order to im- prove the accuracy of the results ,this paper applies the K-means method to cluster the project and calculate the users＇ similarity based on similar projects,proposes an app recommendation combing item features and social networks. The experimental results show that, the app recommendation combing item features and social networks can relieve the influence of the sparsity of user-item ratings on collaborative filtering recommender systems. To a certain extent,it can improve the accuracy of the recommender systems.