Method and system for classifying advertising creativity

Abstract

An embodiment of the present invention provides a method and system for classifying advertisement creativity, including obtaining creative copywriting and creative pictures of advertisement creatives to be classified; using a convolutional neural network to perform feature extraction on the creative copywriting to determine corresponding text features; The product neural network is used to extract the features of the creative picture to determine the corresponding picture features; the text features and picture features are spliced ​​together to obtain advanced features; according to the advanced features, determine the creative features of the advertisement to be classified industry category. The present invention carries out multi-mode document modeling based on creative copywriting and creative pictures, takes creative copywriting and creative pictures as a reference, uses convolutional neural network for feature extraction, obtains text features and picture features, and determines the splicing of text features and picture features Combined together, high-level features are obtained, and then classified according to the high-level features, the industry category of the advertising creative to be classified can be obtained, and the accuracy of classification can be improved.

Description

technical field [0001] The invention relates to the technical field of Internet advertisements, in particular to a method and system for classifying advertisement creativity. Background technique [0002] Programmatic advertising is already one of the important sources of income for many large Internet companies. When advertisers place advertisements on the advertising platform, they need to fill in information such as the industry to which the advertisement belongs, and the advertising platform will conduct special reviews of advertisements in specific industries (for example, whether there are false publicity, vulgarity, etc.). In order to avoid review, some advertisers will deliberately fill in false industry information. Manually identifying whether the industry information filled in by advertisers has obvious disadvantages (high labor costs and poor real-time performance), so it is necessary to have an automated advertising creative classification scheme to identify re...

AI Technical Summary

Problems solved by technology

The problem is that the classification of advertising creatives based solely on copywriting or pictures does not use all the information of this creative, which is equivalent to losing some features. Due to the different emphasis of information conveyed by text and pictures, the accuracy of classification Difference

[0006] 2) Most of the existing classification models are based on statistical methods

[0008] The existing advertising creative classification methods cannot effectively combine the copywriting and matching pictures in the advertising creative, making the classification accuracy relatively low

Images