Research And Implementation Of Personalized Recommendation Of TV Programs Based On Data Mining

As a part of intelligent information service,Personalized Recommendation system is concerned and researched by many industries,and even become the core of some industries.But the particularity of the TV program makes it not widely used.Personalized Recommendation of TV programs becomes the urgent demand of TV service providers and users.This paper focuses on the field of recommendation system,using a variety of programming languages to make Personalized Recommendation for TV programs.In the first half of this paper,we introduce the present situation of Personalized Recommendation of TV programs,and analyze the similarities and differences between it and others.After that,we introduce several classical algorithms of recommendation system and discuss the advantages and disadvantages of each.The emphasis is that we introduce two kinds of Collaborative Filtering and explain them.And then,we introduce the algorithm evaluation index of recommendation system.These works lay a theoretical foundation for later data analysis,Data Mining and specific recommendation of TV programs.In the first half of this paper,we implemented the recommendation system.Including three aspects:(1)Data collection and cleaning.This paper uses Python's Web Crawler technology to collect the types of TV programs,and uses R language,SQL statements to clean the data and building an implicit scoring system;(2)data analysis and mining.This paper uses R language to visualize the sorted data,analyze and mine the useful information hidden in the data.(3)Implementation of recommendation system.According to the principle of Collaborative Filtering algorithm,we use Python language to program.And then,on the basis of the information obtained from the data analysis and mining,the TV programs and program types are recommended to a certain user.Through the implementation of two kinds of recommendation,the article draws the conclusion that the type of recommendation program can be used as the basis for recommendation of TV programs.Compared with other articles,this discovery is not only a conclusion but also an innovation.It also finds that the sparse matrix in data sets leads to the low accuracy of recommendation,which is a problem that needs to be solved in future study.