How to recommend iptv programs

Abstract

The invention discloses an IPTV program recommendation method. The method comprises the steps of building a user rating matrix according to a watching log of a user; sequentially performing low-rank denoising, periodic compression and watching mode extraction on the user rating matrix, so as to obtain a periodic watching scene of the user; using each watching mode in the periodic watching scene as an interest user, and building a score matrix of all interest users; and identifying a current watching mode so as to determine a current interest user, recommending programs that the user is interested in from a historical film bank and current live programs by using a joint recommendation policy, generating a recommendation list and pushing the recommendation list to the user. The method can well handle the multi-member condition at an IPTV terminal, improving the recommendation precision. In addition, a live and on-demand combined recommendation policy is adopted, so that the user can acquire more comprehensive recommended content.

Description

technical field [0001] The invention relates to the technical field of recommendation, in particular to a method for recommending an IPTV program. Background technique [0002] With the rapid development of the Internet, Internet TV has become more and more popular. There is usually a single member behind the Internet terminal, so the recommendations on the Internet are all for the interests of a single member, and only those who are interested can be recommended. However, IPTV has its own particularity. There are multiple members behind each IPTV terminal. Therefore, applying the recommended method in the Internet field to the IPTV field will greatly reduce the effect. Although there are already some recommendation algorithms on IPTV, none of them make recommendations by distinguishing multiple members. Since there are multiple members behind the IPTV terminal, the interest of the IPTV terminal is the coupling of the interests of the members behind it, but each member has...

AI Technical Summary

Problems solved by technology

The above-mentioned patent directly recommends programs that the user is interested in based on the user's current interests, without considering the similarity between users or between programs, making it difficult to discover the potential interests of the user, making the recommended content limited

In addition, when recommending specific programs, most personalized recommendations based on collaborative filtering are only processed for on-demand programs, ignoring most of the live broadcast behaviors of radio and television users.

Images