Research on Video Recommendation Algorithm Based on Hybrid Collaborative Filtering

This paper proposes a hybrid collaborative filtering algorithm for video recommendation, based on user profile, item characteristics, and time weighting. This approach addresses user bias inherent in traditional collaborative filtering algorithms, enhancing recommendation personalization and accuracy. Leveraging this algorithm, a video recommendation system is implemented, utilizing Spark Streaming's distributed computing capabilities to improve processing speed and scalability. Experiments conducted on video datasets verify the algorithm's efficiency and effectiveness. The proposed method demonstrates broad applicability and practical significance.