Drugs Classification Full Project

This project develops a drug classification model to predict the most suitable medication for patients based on age, sex, blood pressure, cholesterol, and sodium-to-potassium ratio (Na/K).

Workflow:

Exploratory Analysis: Middle-aged patients dominate the dataset. High BP and cholesterol are strongly linked to DrugY, while high Na/K ratios show a strong influence on prescriptions.

Preprocessing: Encoded categorical variables and normalized continuous features.

Modeling: Applied Decision Tree, KNN, Naive Bayes, Random Forest, and Logistic Regression, evaluated with accuracy, recall, F1-score, and confusion matrices.

Results: Best-performing model achieved high accuracy. Feature importance revealed Na/K ratio, BP, and cholesterol as the key predictors.

Key Insights:

Middle-aged patients are overrepresented, influencing drug prescription trends.

DrugY is most common among patients with high BP, high cholesterol, and elevated Na/K.

The final model serves as a reliable decision-support tool for doctors in prescribing drugs.