Research And Implementation Of Radio And TV Program Recommendation System Based On Knowledge Graph

As technology constantly develops,it brings more convenience for people's lives,which also makes their amateur lives become more colorful.As for the main way adopted by people to obtain film and television programs at present,it has tuned to Internet platforms from television channels.With more choices,the time people uses to watch programs is more flexible.Under this situation,it can be seen that everyone is enjoying the convenience brought by the development of information.However,the problem of information overload is also caused with the development of information technology.People often have difficulty in quickly choosing their favorite content in the face of a large number of film and television programs on the Internet.In this context,it brings the emergence of personalized recommendation system,which solves the problem of information overload to a certain extent,so as to make users able to find the items they really like faster.On the other hand,more and more valuable recommendation models have been proposed as the deep research on the field of personalized recommendation is carried out by many scholars.Thus,some of the problems of traditional recommendation models are alleviated by these new recommendation models to a certain extent,such as the problems of data sparseness and cold start.More scholars began to carry out the research on the concept of knowledge graph since it was proposed by Google in 2012.When it comes to knowledge graph,it has also shown great achievements in the field of personalized recommendation,which has promoted the generation of many valuable recommendation models based on knowledge graph.In line with the recommendation model of the knowledge graph,this thesis takes advantage of actual production environment data and Internet data to make the construction of a broadcasting and TV program recommendation system based on the knowledge graph,which also combines with the actual project scenarios.As for the broadcasting programs mentioned here,it mainly covers films,TV series and animation programme in the radio and television industry.This thesis completely presents the whole process of building the system from the requirements analysis to the system design,and then to the implementation and testing of the system.What's more,the knowledge graph in the field of film and television is constructed in this thesis,and a new recommendation model KBT based on knowledge graph is adopted.A better recommendation effect is achieved by comparing it with the traditional film and television program recommendation system.And the following is the main work of this thesis:1.The film and television program information in the Internet and the user behavior interaction records in the actual production environment are collected and handled;2.The knowledge graph of film and television programs in the field of radio and television is constructed,which contains more than 80000 films,TV dramas and animation programs;3.The experiment on recommendation models based on knowledge graphs is carried out,which filters out the best recommendation models to apply them to the radio and television program recommendation system that we want to build based on knowledge graph;4.By taking the knowledge graph constructed in the second step as the data basis,it designs and develops the recommendation system of radio and television programs based on knowledge graph by combining with the improved recommendation model in the third step,which also carries out the test on the system.Furthermore,most of the calculation work in the offline layer is carried out,which can guarantee the good user experience and system performance,so as to meet the real-time demand of users.