Evaluation and Improvement of Media Personalised Recommendation Effect Based on Collaborative Filtering

With the explosive growth of digital media content, personalized recommendation is indispensable. Among them, collaborative filtering is a widely used algorithm with the characteristics of accuracy and simplicity. This study aims to evaluate personalized media recommendation based on collaborative filtering and explore its potential improvement strategies. We first analyzed the application of existing collaborative filtering methods in media recommendation, and then evaluated their performance on different datasets. In addition, we also explored in depth how to improve the algorithm effect, including improving the diversity and novelty of recommendations. The recommendation effect can be significantly improved by combining strategies such as hybrid models and considering timeliness and user behavior dynamics. The content of this paper provides new insights and methods for personalized recommendation in the media field.