基于物品内容的电视产品推荐系统研究

Research on TV Product Recommendation System Based on Item Content

网络电视资源量的不断膨胀,为用户寻找资源加大了难度。此时对用户的电视产品推荐功能就显得尤为重要,如何对用户做出精准的推荐服务成为一个亟待解决的问题。针对现有的一些方法在冷启动、新项目等问题上的不足,论文利用了某广电网络运营公司的38010条收视行为数据和18480条节目信息数据,通过分析两者的内容相似度,构建了一种基于物品内容的电视节目推荐系统,并通过准确率和召回率两个评价指标对系统进行了评价。实验结果表明该方法的准确率和召回率到达了25%和14%以上,能够将新节目准确地推荐给用户。

The continuous expansion of the amount of Internet TV resources has increased the difficulty for users to find resources. At this time,the recommendation function of users’ TV products is particularly important. How to make precise recommendation service to users has become an urgent problem. In view of the shortcomings of some existing methods in cold start and new projects,this paper uses 38010 viewing behavior data and 18480 program information data of a radio network operation company.By analyzing the content similarity of the two,a TV program recommendation system based on item content is constructed,and the accuracy rate is passed through the accuracy rate. And the system is evaluated through two evaluation indicators of accuracy and recall. The experimental results show that the accuracy and recall rate of the method reach 25% and 14%,which can accurately recommend new programs to users.