基于协同过滤算法的短视频平台个性化推荐模式研究

Research on Personalized Recommendation Mode of Short Video Platform Based on Collaborative Filtering Algorithm

短视频平台主要通过短视频的个性化推荐,提高用户定点投放能力,但短视频平台个性化推荐的错误率高,因此,提出基于协同过滤算法的短视频平台个性化推荐模型。在缓存域内,对短视频平台个性化源数据进行自适应统计特征分析,提取短视频平台个性化特征参数,用联合关联特征分析方法,计算短视频平台个性化参数,采用标签化控制方法,结合用户对相关资源的预测偏好,采用协同过滤算法,实现用户的兴趣标签分类和资源偏好识别。根据评级数据和标签信息定义结果,实现短视频平台个性化推荐。实验结果分析得出,该方法进行短视频平台个性化推荐的错误率较低,且用户满意度较高,在最优状态下推荐的满意度均值为84.68%。

Short video platform mainly improves users’ ability of targeted delivery through personalized recommendation of short video, but the error rate of personalized recommendation of short video platform is high. Therefore, a personalized recommendation model of short video platform based on collaborative filtering algorithm is proposed. In the cache domain, the personalized source data of the short video platform is analyzed by adaptive statistical features, the personalized feature parameters of the short video platform are extracted, the personalized parameters of the short video platform are calculated by the method of joint correlation feature analysis, and the user’s interest tag classification and resource preference identification are realized by the tag control method combined with the user’s prediction preference for related resources and the collaborative filtering algorithm. According to the rating data and tag information definition results, personalized recommendation of short video platform is realized. The experimental results show that the error rate of personalized recommendation for short video platform by this method is low, and the user satisfaction is high, the average satisfaction of recommendation in the optimal state is 84.68%.