case explore (by 3.24)

[wangqianwen0418]

:mag:case exploration a new potential dataset, student performance https://archive.ics.uci.edu/ml/datasets/student+performance

[wangqianwen0418]

Dataset: https://www.kaggle.com/aljarah/xAPI-Edu-Data

• sample_num=1000, sup_min=5; positive risk=决策利于女性
• origin data: existing discrimination to both female and male, risk: [ -0.3, 0.5]
• xgboost: discrminination to males only, risk: [0, 0.23]
• logistic regression: discrminination to males only, risk [0, 0.33]

[wangqianwen0418]

• Academic for one model;
• Adult dataset for comparing models;

[wangqianwen0418]

knn: original model

knn_post_1: using the rejection option (a post-processing method) to remove algorithmic discrimination

knn_post_2: rejection option + observertion from DiscriLens