目录

Implicit User Awareness Modeling via Candidate Items for CTR Prediction in Search Ads

This repo is the official implementation for the SIGIR 2023 paper: LOVF: Layered Organic View Fusion for Click-through Rate Prediction in Online Advertising.

Data format

Our released dataset ORAD-Pub could be found at JD JingPan with password 3jhbd6.

In the data files, each row corresponds to a search session. Each column is a piece of multiple sample data aggregated according to user-query. The organization form of each column is:
column[0]: user_id. int type
column[1]: query_id. int type
column[2]: The source of each sample(0:advertising. 1:organic). list type
column[3]: The label of each sample. list type
column[4:]: the side info [itemID, categoryID, brandID, vendorID and priceID] of each sample. list type

Requirements

  • python 3.6.13
  • tensorflow 1.15.0
  • scikit-learn 0.24.2

Quick start

Create a new data folder and put the downloaded dataset into the folder. Then,

python src/main.py 
关于
41.0 KB
邀请码
    Gitlink(确实开源)
  • 加入我们
  • 官网邮箱:gitlink@ccf.org.cn
  • QQ群
  • QQ群
  • 公众号
  • 公众号

版权所有:中国计算机学会技术支持:开源发展技术委员会
京ICP备13000930号-9 京公网安备 11010802032778号