An implementation of a data generation process for treatment effect estimation from [Powers, 2018]
[Powers, 2018] Scott Powers, Junyang Qian, Kenneth Jung, Alejandro Schuler, Nigam H. Shah, Trevor Hastie, Robert Tibshirani. (2018). "Some methods for heterogeneous treatment effect estimation in high-dimensions." Statics in Medicine, 37(11): 1767--1787.
from generator import generate_data
# Select scenario, the number of data (patients) and the number of features for generating data.
x, t, y, treatment_response, control_response = generate_data(4, 200, 100)
# True CATE can be obtained by substracting control_response from treatment_response.
true_conditional_treatment_effect = treatment_response - control_response