TV Ratings Prediction Research Based On Decision Tree Algorithm

In the modern television media operators,since the market competition mechanism and gradually improve the economy,government funding support will be less and less,relying on limited resources and therefore how to add revenue,it is essential for television media.Television ratings is a very important indicator,not only related to the television channel reputation among the masses but also on television itself infinite operational benefits;and the need to purchase advertising time on television business customers,television ratings rate irregularities also directly affect the price of purchase advertising time irregularities,those great expectations for companies to promote their products high television ratings will no doubt bring a very good role not competitive channel is undoubtedly only spoil the company's time and money,or the company will select some important brands into these channels with low viewership of broadcast.In today's new media operations,an effective prediction for television ratings is a problem,is also a need we keep learning course.There are many factors can affect the rating of TV programs,we need to study how to take these factors into our research related index in the model,at the same time in the expression in relatively accurate digital display,this will weaken because of the error of the subjective factors bring to predict television ratings.In order to improve the accuracy of the prediction,this paper first to get the preliminary processing of raw data,data clearly some interference.Then this paper studies how to establish and realize audience rating system.First,Introduce the audience rating's definition,effect and investigating method in detail,Analyze the predicting mathematic model and establish 10 main influence factors.The experiments indicate that the effect is well.At first this paper studies how to combine the 10 main influence factors to establish and realize audience rating system using Artificial Neural Network and Decision Trees technology.Secondly this paper studies how to combine the 10 main influence factors and 3 subjective factors to establish and realize audience rating system using Bayesian Network technology.The experiments indicate that the Decision Trees is better than Artificial Neural Network and Bayesian Network in predicting TV audience rating,because combining 3 subjective factors.