Model Research And System Implementation Of Video Evaluation Based On Big Data

In recent years,with the continuous development of China's network technology and the "triple play",IPTV(Internet Protocol Television)showed a rapid development trend.IPTV is different from the traditional one-way television broadcast mode.It can achieve the interaction with the audience,and the videos are more personalized and diversified.At the end of June 2016,the number of IPTV users in China has reached 65.81 million.Therefore mining and meeting the needs of users has become the key of IPTV development.How to evaluate a large number of videos and choose a more popular video has become an urgent problem to be solved in IPTV.With the rapid development of new media,the massive video resources and a large number of video related information arrived,the video is no longer confined by the video ratings of traditional platform,the ratings in the new media platform,network effects and the video itself also affect the evaluation of the video.This paper uses the massive data of new media platform to improve the IPTV video evaluation index system,and constructs video evaluation model on the Spark platform based on the BP neural network.Video evaluation system was set up to facilitate the video evaluation of IPTV.The main contents and innovations of this paper are listed as follows:Firstly,combined with massive video information,this paper improves the video evaluation index system from the video rating,video network impact and the video itself and points out the various indicators of data sources and quantitative criteria.Secondly,this paper introduces the video implicit score,and constructs the video evaluation model based on BP neural network.The video evaluation index is used as input data and video implicit score as output data.The experiment shows that the constructed evaluation model can evaluate the video more comprehensively and effectively.In order to improve the efficiency of constructing the evaluation model,this paper conducts the BP neural network parallel training on the Spark data processing platform.The experiment proves that it can effectively improve the efficiency of the evaluation model.Finally,based on the video evaluation index system and the video evaluation model,a video evaluation system a video evaluation system is established.The system collects the data related to each index in the video evaluation index system and integrates and supplements the data at first,and then embeds the weights and thresholds of the video evaluation model into the system,which facilitates the evaluation of the video.