zhaoxianglin / 7160project

Comp7160 Project

2022 ZHAO Xianglin

The results of this experiment based on the "criteo_sample_50w" dataset. Dataset DownLoad: https://lifehkbueduhk-my.sharepoint.com/:x:/g/personal/22481079_life_hkbu_edu_hk/EaL8f9cxxx9OhIGMejbeT-MBLRmz0HBgVsAx-Fc0WFwZBA?e=S8ruNd

The original dataset which can be used to train this model: Dataset Download: Display Advertising Challenge https://www.kaggle.com/c/criteo-display-ad-challenge

Run the model

Step 1:

pip install -r requirements.txt

Step 2: Download the dataset and put the "criteo_sample_50w.csv" file into the data folder.

Step 3: Open the test_model.ipynb and run each cell. You can see the results of the experiment directly in this file.

Reference:

DeepFM: A Factorization-Machine based Neural Network for CTR Prediction https://github.com/chenxijun1029/DeepFM_with_PyTorch https://github.com/SpringtoString/DeepFM_torch