基于数据挖掘的视频推荐系统研究

北京理工大学珠海学院2020届本科生毕业论文Research on video recommendation system based on Data MiningAbstractThis paper investigates how to efficiently and accurately recommend videos of interest tousers based on a personalized recommendation system based on a large amount of behav-ioral data of users.The study is based on the following three main components.First,the SVD(Singular Value Decomposition)algorithm is used to score predictions forvideos that are not rated by users based on video similarity.The ALS (Alternating Least Squares)algorithm is then used to rate predictions for videosthat are not rated by users based on their liking of the tags contained within the video.Finally,descriptive statistics based on prevalence make popular recommendations.Basedon the SVD algorithm,the model is trained by the ALS al gorithm and validated on the dataset to obtain a more accurate and personalized recommendation system.Keywords:Recommendation system Collaborative filtering SVD ALS