Online advertisement audience sorting method based on transfer learning

An online advertising and transfer learning technology, applied in the field of Internet advertising data preprocessing, can solve problems such as not having potential interest intentions, achieve high advertising click-through rate, improve the effect of good results, and reduce the impact of non-related queries

Inactive Publication Date: 2015-03-04
BEIHANG UNIV
View PDF4 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although a user's historical behavior has a strong correlation with their potential interest in advertisements, d

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Online advertisement audience sorting method based on transfer learning
  • Online advertisement audience sorting method based on transfer learning
  • Online advertisement audience sorting method based on transfer learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] The overall flow chart of the online advertising audience ranking method based on transfer learning of the present invention is as follows figure 1 shown. The overall process includes four parts: data preprocessing, feature extraction, model training, and effect evaluation.

[0025] (1) Data preprocessing

[0026] 1. Extract advertisement title and description information

[0027] Online advertisements usually provide ad titles and ad descriptions to present the specific content of the ad, and the ad description is a more detailed expression of the ad content than the ad title. The advertisement title belongs to concise short text information, and the advertisement description belongs to detailed long text information. This method represents an online advertisement by extracting and segmenting the advertisement title and description information, and using the bag-of-words model in the vector space model.

[0028] 2. From Internet historical logs, extract users’ long...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an online advertisement audience sorting method based on transfer learning and aims at sorting audiences according to potential correlation of Internet audiences about an advertisement, so that a long-tail advertiser with insufficient budget can put the advertisement through buying a fixed number of front sorting user positions. According to the method, a history behavior domain and a preference domain of the advertisement of a user are in different characteristic spaces, and the relation between the history behavior domain and the preference domain of the advertisement of the user is established through defining a correlation transfer matrix, so that the influence of history behavior information of the non-correlated users on advertisement preference is reduced, and thus an improved correlation computational algorithm TransferBM25 based on transfer learning is realized. On the basis, through four steps of preprocessing, feature extraction, model training and effect evaluation on the advertisement history data, a final audient sorting model is obtained, and the advertisement audiences are sorted according to the model, so that after being sorted, the users with the front sorting positions have higher advertisement click probability.

Description

technical field [0001] The present invention relates to Internet advertisement data preprocessing, feature extraction, model training, and effect evaluation methods, in particular to an online advertisement audience sorting method based on migration learning. Background technique [0002] The rapid development of technology and the Internet in the world has driven a new industry with great economic value - online advertising industry. Different from the traditional offline advertising delivery model, online advertising is interactive, customizable, trackable, and deliverable. According to its characteristics, in order to optimize the effect of online advertising and maximize the benefits of advertisers, media, and audiences, a new discipline "computational advertising" was born. "Computational advertising" aims to obtain the most matching advertisement based on the given user and contextual content through calculation and carry out precise targeted delivery. [0003] Behav...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F17/30
CPCG06F16/9535G06Q30/0251
Inventor 张立鑫陈真勇陈朋杰熊璋
Owner BEIHANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap