Research On Video Content Quality Evaluation And Recommendation Algorithm Based On User Behavior

With the rapid development of the Internet,online learning has become one of the important ways for people to learn.The online learning video website provides users with massive teaching video resources.The recommendation algorithm can help users quickly find suitable resources from massive video resources,improve user experience,and increase user stickiness and loyalty.Therefore,the recommendation algorithm has gradually become a hot research topic due to its wide application.In the traditional video recommendation algorithm,the inefficient description of the user's video watching process leads to low recommendation efficiency.Based on the collaborative filtering recommendation algorithm,this paper proposes a video content quality evaluation and recommendation algorithm based on user behavior.First,collect the record and time point of the user's behavior in the process of watching the video,extract the behavior characteristics of the user on the watched video;then,construct a neural network and combine the fuzzy comprehensive evaluation method to comprehensively analyze the user's attitude to the watched video The user's interest tendency;Finally,the user's interest tendency is clustered and similarity evaluated based on the user's attitude and interest tendency of the video,and Top-N video recommendation is made according to the evaluation result.The research contents of this article are mainly as follows:(1)Research the relationship between user behavior characteristics and personal interests in the viewing process,and proposes a description method of user viewing process based on FBP(Fast Backward Partial).The method analyzes the user's behavior and time point during the viewing process,and obtains the number of fast forward rewind and playback time ratios,which is used to analyze the relationship between the user's behavior characteristics and personal interests.(2)Aiming at the lack of behavior analysis and objective evaluation of video content quality during the viewing process of video websites,a video content quality evaluation algorithm based on user behavior is proposed.The algorithm first collects the user's behavior and time during watching the video,and uses the FBP model to model the user.Then construct a neural network model for analyzing the user's attitude,and obtain the user's viewing attitude.Finally,the video content quality is evaluated objectively based on the attitude of all users watching the video.(3)Aiming at the problem that the traditional video websites do not detailly describe the user's viewing process,judge of user interest unclearly and inaccurate recommendation,a personalized video recommendation algorithm based on user behavior is proposed.The algorithm first preprocesses user behavior data and uses the FBP model to model the user.Secondly,the user's attitude is judged by the fuzzy comprehensive evaluation method,and the user's interest and viewing intention are analyzed.Then users are clustered through the C&S model,and the user similarity is calculated using the Jaccard similarity,and videos that do not overlap between similar users are searched for Top-N recommendation.Finally,build a teaching video website to collect user behavior data to verify the effectiveness of the video recommendation algorithm based on user behavior.Experimental results show that,under the same evaluation index,the algorithm in this paper is superior to the traditional video recommendation algorithm and has certain practical value.