Design And Implementation Of Video Service Platform Based On Content Recommendation

With the development of science and technology,Internet technology has gradually penetrated into all walks of life,and online video has greatly enriched our lives.However,due to the influx of a large number of video information,users cannot get the video information they want accurately and quickly,which leads to a decline in user experience.Therefore,in order to solve users' pain points,this paper provides personalized recommendation service for users through hybrid recommendation algorithm.The system adopts the structure system of B/S mode,based on the MVC design mode,and SSH framework(Struts+Spring+Hibernate).The system USES JSP technology in the page,MySQL as the database management system,and comprehensive utilization of data through Web Service.The system mainly includes the following five functional modules:(1)User management module:mainly responsible for the management of user information,complete the functions of logging in,registering and viewing the personal homepage,and at the same time,display the user's playing history,attention information and favorite information.(2)Work management module:the work management module is mainly divided into two parts,including users' management of video uploading,editing and deletion;It also includes the operation of creating,editing and deleting the broadcast list;At the same time,the video can be added and deleted in the playlist.(3)Search module:the search module can mainly display users' historical search records and the hot search words in the system,and can also conduct fuzzy query through keywords.(4)Video details module:users can click on a video to enter the video details page,where they can control the video playing and also follow,thumb up and comment on the video.(5)Video recommendation module:through the user's historical behavior data,the mixed recommendation algorithm is used to recommend the videos that the user may like,so as to explore and expand the user's interest and preference and at the same time filter the videos that the user does not like.Through the comparison and analysis of different recommendation algorithms,combined with the advantages of content-based recommendation and user-based collaborative filtering recommendation,the hybrid recommendation algorithm is designed and described.The acquired data are cleaned up,and the user video matrix is constructed to record the user's behavior towards the video.The video information matrix is constructed by extracting the feature values from the video text information,so as to establish the user interest model and filter the videos disliked by users,so as to complete the recommended behavior,improve the user experience and facilitate users to find the required videos accurately and quickly.