midterm report peer review #92
Description
This report is about improving algorithmic fairness in COMPAS. I do think the author should elaborate a little more on explaining what is COMPAS and what is the historical bias behind it.
In the data preliminary analysis part, I really like the histograms. I think it's very simple and clear to reveal some insights of the preliminary data.
The report used logistical regression to fit the initial data which I think suit the situation very well. The False Positive and False Negative rate indicate there are bias in the COMPAS algorithm towards white and black defendants.
The future plan looks promising. Overall I think this is a clear and concise report.