Research And Application Of Video Ratings Analysis Algorithm Based On Big Data

Audience rating is a very important indicator of the television program,which has important significance for the management of TV station and precision advertising.However,today's ratings survey is not perfect,quantitative analysis cannot be done on single ratings survey data,ratings cannot be predicted before the show plays in the system in which the investigation,analysis and evaluation are conducted after broadcast.Therefore,modification of program type and adjustment of new program type relies on personal subjective experience instead of science quantitative research.If expected broadcast television rating is predicted,we can schedule programs better.This paper studied on how to predict TV ratings effectively,introduced the research background and significance of TV ratings prediction,sulmmed up a series of data pre-processing scheme according to the characteristics of video data,put forward a prediction scheme of film and television program with naive bayes classification based on hadoop and got acceptable results,verified the feasibility of expansion and stability of performance in the distributed naive bayes algorithm in rating prediction through analysis on matlab and hadoop and provided theoretical basis for the TV stations to optimize program arrangement and accurate advertising.