Advertisement content recognition method and device

Abstract

The invention provides an advertisement content recognition method and device. The advertisement content recognition method comprises the following steps: scanning titles and main bodies of articles published by a user in a community application; when scanning key information, judging whether the key information belongs to a corresponding key information blacklist set or not; if the key information does not belong to the corresponding key information blacklist set, carrying out global article searching by using the key information; and when the emerging frequency of the key information in global articles is greater than a preset threshold value, recognizing the articles as advertisement articles, and shielding the advertisement articles. The advertisement articles are automatically recognized according to characteristics of the advertisement articles, repeated complicated sample learning is not required, so that maintenance cost of a system can be reduced, the problem of hysteresis of treatment to variant advertisement articles by a machine algorithm or a method of complaints from customers and the like can be solved effectively, manual intervention check is seldom required, and running cost can be greatly reduced.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a method and device for identifying advertisement content. Background technique [0002] With the increasing abundance of Internet information, the traffic of posting and viewing posts in forums and blogs is increasing. Many unscrupulous users try to earn traffic from forums and blogs by repeatedly posting some advertising content. It poses a great threat to the user experience and seriously affects the browsing and use of normal users. [0003] The following methods are commonly used to identify advertising content: [0004] 1) Through the machine learning algorithm, machine learning is performed on some previous advertisement samples, and then the posts are scored to further identify the advertisement posts; [0005] 2) Through user complaints and reports, the operator will judge the post afterwards; [0006] 3) All posts are allowed to go online after being manually re...

AI Technical Summary

Problems solved by technology

Through this method, according to the characteristics of the advertisement posts, the advertisement posts are automatically identified without repeated complex sample learning, which can reduce the system maintenance cost, and can effectively solve the problem of machine algorithms or rely on user complaints to deal with the flood of advertisement post variants. Hysteresis issues, less need for manual intervention review, which can greatly reduce operating costs

Images