基于特征偏好分析的改进混合推荐算法

Improved Recommendation Algorithm Based on Latent Reviews Analysis

摘要: 随着互联网步入大数据时代，网络上的海量数据为人们提供极大便利的同时，与之相伴而来的是信息过载问题。当前协同过滤算法是解决信息过载的主流推荐算法，但传统推荐算法面临着数据矩阵稀疏性问题和冷启动问题，从而影响个性化推荐的准确性。本文主要研究的是基于特征偏好分析的改进混合推荐算法，该方法将分析用户特征偏好和物品特征相结合，再使用传统的协同过滤思想，将最优评分对象推荐给用户。实验表明，该算法有效地提高了推荐结果的准确性。

Abstract: With the coming of big data era, massive data on the Internet provide people with great conven-ience. But at the same time, the problem of information overloading becomes more and more obvious. At present, collaborative filtering algorithm is the mainstream recommendation algorithm to solve the problem of information overloading. However, the traditional recommendation algorithm is faced with the problem of data matrix sparsity and the problem of cold start. This paper proposes a hybrid recommendation algorithm based on features analysis and combines the idea of collaborative filtering and latent semantic analysis. Finally it recommends the optimal objects to users. The experimental results show that the improved hybrid recommendation algorithm can improve the accuracy of the recommendation results.