Crowd mining system for TV program based on audience behavior analysis

This paper studied the information overload brought by abundant digital television (TV) program resources and media image which need to adapt to the changing market environment by crowd mining based on audience behavior analysis. When the audience crowd is classified to several levels, personalized audience behavior analysis method and group audience behavior analysis method are proposed separately. It is pointed that for proposed crowd mining system, data mining algorithm was used to analyze the data through audience characteristic and viewing effect, then, the actual audience distribution was obtained by demographic features. The results indicate that it could help the decision maker managing the program contend and broadcast time seasonably.