How to generate the Fig.2a?

[horacehht]

I am primarily focused on a specific PR-curve. PPI presents a multi-label problem, and to the best of my knowledge, it's standard to draw a PR-curve for each class. However, in Fig.2a, a model is represented by a single line. How is the transition made from a multi-label PR-curve to a single curve?

[horacehht]

I have encountered a challenge while working with PR (Precision-Recall) curves. For each threshold value, there is a corresponding precision and recall, which I have extracted using the sklearn.metrics.precision_recall_curve function. My initial plan was to calculate the average precision across various thresholds. However, I observed that the lengths of the returned precision arrays are inconsistent, rendering the averaging process unfeasible. Subsequently, I sought guidance from an AI model (GPT), which suggested interpolation as a potential solution. Acting on this advice, I proceeded to generate a curve through interpolation. Unfortunately, the resulting curve appears anomalous and does not align with the representations shown in your publication.