Analysis of TV audience flow based on the lingered channel matrix

At present, the analysis of television ratings gives priority to ordinary sectional analysis which is represented by audience ratings, market share and other indices, and has little analysis of the audience flowing. Additionally, traditional research method of audience flowing has some problems such as high time cost, large storage space, weak reusability and no analysis about overall flow trend or audience clustering. This paper proposes a new scheme on the analysis of audience flowing. It is based on the established lingered channel matrix produced by raw viewing data. From the angle of channels (or programs) and users, this paper adopts the conventional mathematical statistics method and clustering algorithm to reflect audience flowing between different channels and their own viewing behavior in detail. The lingered channel matrix established by the scheme clearly shows the channel that audience linger at every moment, leaving the original viewing data away and avoiding the problem of high cost of time and storage space by repeatedly reading and processing raw data. Moreover, the matrix provides the data base for the subsequent extension research, in order to realize the analysis of TV audience flowing in a deeply vertical level.