############################ Environment: python 3.7
############################
python main_realworld.py/main_synthetic.py int(times)
a) python main_realworld.py 100 Input: realworld.npy (realworld dataset) Output: result1_realworld.csv emp1_realworld.csv
b) python main_synthetic.py 100 Output: synthetic_gaussian.npy (synthetic dataset with Gaussian errors) result1_synthetic_gaussian.csv (statistics) synthetic_cauchy.npy (synthetic dataset with Cauchy errors) result1_synthetic_cauchy.csv (statistics) emp1_synthetic.csv (all empirical errors)
% Due to privacy issue, we do not provide the real-world dataset. You can change the input dataset to be your own panel dataset "XXX.npy" in which the ((iT-T+t)-th row is (x_{it}, y_{it}), i.e., the observation of i-th individual at t-th time period.
############################