Questions about the Metrictor_PPI function

[wan-nie]

I would like to express my gratitude for making the code open-source. I have a question regarding the Metrictor_PPI function in HIGH-PPI/utils.py and I was hoping you could assist me.

• It seems that the arguments of the sklearn precision_recall_curve function are reversed. Could you kindly clarify if this is the case?
• self.pre is (0,1) label instead of the predicted probability $P(y=1)$. Line 150 in HIGH-PPI/model_train.py. I am concerned that this could affect the correctness of AURPC.

[zqgao22]

Hi, Wan. Thank you very much for pointing this issue. You are right. For Sklearn's precision_recall_curve function, we need to take the predicted probability as the second entry. In fact, the result of our source data is calculated with a simpler function average_precision_score, which also requires the predicted probability scores. However, we later tried to rewrite it to a more widely used precision_recall_curve function and ignored the above problem. We are about to modify it to the correct form.

[wan-nie]

Hi, Ziqi. Thank you for the prompt reply. As the results were correctly calculated using the predicted probability, I think this issue can be closed now.