RecommendationAlgorithmPractice

This program is the implemention of "The Theory and Practice of Recommendation Algorithm". It mainly contains feature enginnering and model implemention.

本项目为 "推荐算法和实践" 的代码教程，主要包括特征工程和模型实现。

The link of "The Theory and Practice of Recommendation Algorithm" is published in 推荐算法的理论和实践

理论讲解发布在推荐算法的理论和实践

Note:

The dataset of this program includes the income census, BX-CSV-Dump.

The model is implemented by tensorflow1.0

注：

本项目的数据集包括income census，BX-CSV-Dump

模型代码实现均为tensorflow1.0

Feature Enginnering(特征工程)

Continuous data(连续数值处理)

• Data Standardisation(数据标准化)
• Continuous data binning(无监督分箱(unsupervised binning),IV分箱(IV binning),卡方分箱(chi2 binning),KS分箱(Best KS binning)和最优分箱(optimal binning))

Discrete Data(离散数值处理)

• Label Encoding
• Frequency Encoding
• OneHot Encoding
• Target Encoding
• CatBoost Encoding
• WOE Encoding
• MEstimate Encoding
• LeaveOneOut Encoding
• Sum Encoding
• BetaTarget Encoding

Collaborative Filtering(协同过滤)

• Item Based CF
• User Based CF
• SVD
• Swing
• Adamic-Adar