You are a computational medicinal chemist working on a small-molecule drug discovery program targeting a clinically relevant enzyme implicated in human disease.
Experimental screening has already been performed on a subset of compounds, generating quantitative bioactivity data (pIC₅₀). Due to cost and time constraints, it is impossible to experimentally test all available compounds.
Your role is to build robust QSAR regression models to predict compound potency and guide medicinal chemistry decisions.

**Objective**
Build accurate QSAR regression models that predict compound bioactivity and minimize Mean Absolute Error (MAE) on unseen compounds.
MAE is chosen because it directly reflects medicinal chemistry relevance:
•	Lower MAE = smaller error in predicted potency
•	Fewer costly experimental mis-prioritizations

Participants must:
1.	Build a QSAR regression model
2.	Predict pIC₅₀ values for the test compounds
3.	Submit predictions to Kaggle