Time-phased and crowd-phased intelligent television advertisement recommendation system and recommendation method thereof

Abstract

The invention discloses a time-phased and crowd-phased smart television advertisement recommendation system, which classifies and labels crowds and constructs date attributes, and further comprises anadvertisement management platform used for calling advertisement information in a data updating cache of a data cache center after the advertisement information is changed; a media asset operation platform; a big data cleaning program which is used for cleaning the user behavior data and separating out user viewing behavior data; a time period-crowd prediction algorithm model group which operatesat a fixed time every day, and analyzes and processes the user viewing behavior data of the user in each time period of the last day; the data cache center which is used for regularly updating data every day; and a time-phased and crowd-phased advertisement recommendation cloud platform which is used for receiving the request and feeding back advertisement data. The invention further provides a time-phased and crowd-phased intelligent television advertisement recommendation method, the advertisement content is recommended according to the crowd to which the television user belongs during thetime period, and the cost performance of advertisement putting is improved.

Description

technical field [0001] The invention relates to the technical field of smart TVs, in particular to a smart TV advertisement recommendation system and a recommendation method for different time periods and groups of people. Background technique [0002] With the popularization of smart TVs, more and more advertisers are willing to place advertisements on smart TVs. Today’s TV terminal advertisements are either too wide and cost-effective, or are limited by the ratings, the user acceptance rate is not high, or due to There is a fixed time for live TV advertisements. Users change channels when playing advertisements, or when children watch TV, they play advertisements for the elderly, etc., resulting in low conversion rate of advertisements. There are three main categories of existing precise advertisement delivery methods: [0003] 1. By segmenting the basic information of live TV, film and television programs and advertisements, and then using algorithms to perform one-to-man...

AI Technical Summary

Problems solved by technology

[0002] With the popularization of smart TVs, more and more advertisers are willing to place advertisements on smart TVs. Today’s TV terminal advertisements are either too wide and cost-effective, or are limited by the ratings, the user acceptance rate is not high, or due to There is a fixed time for live TV advertisements. Users change channels when playing advertisements, or when children watch TV, they play advertisements for the elderly, etc., resulting in low conversion rate of advertisements. There are three main categories of existing precise advertisement delivery methods:

[0003] 1. By segmenting the basic information of live TV, film and television programs and advertisements, and then using algorithms to perform one-to-many matching to recommend advertisements based on live programs, the disadvantage is that the content of advertisements and the content of live TV programs can actually be related to each other. Large or timely associations are also fuzzy associations, lack of precision in delivery, and need to maintain the association relationship between advertisements and live programs, and at the same time have conflicts of interest with large advertisers;

[0004] 2. Associate the advertisement with the video. When the user plays the video, use the obtained video information to query the advertisement associated with the video and push it accurately. The disadvantage is that the advertisement content and the video program content can actually be associated with a small amount Or if it is related in time, it is also fuzzy, the delivery accuracy is lacking, and the relationship between advertisements and video programs needs to be maintained;

[0005] 3. Associate the advertisement with the advertisement, use the user viewing history, analyze the preference score of the advertisement in the history, and recommend similar advertisements. The disadvantage is that although the advertisement that the user may like is obtained, this kind of recommendation will be like an e-commerce recommendation , it is likely to be the same type of advertisement for a period of time, which is likely to arouse user disgust

Images