💉 Analysis of Drug Types for Diabetes Management Using KNN

A machine learning clinical decision-support project using the K-Nearest Neighbors (KNN) classification algorithm to analyze patient health indicators and predict suitable medication types for diabetes management.

---

📌 Project Overview

This project analyzes patient demographic profiles and clinical features (drug200.csv dataset) to recommend optimal drug treatments using K-Nearest Neighbors (KNN.ipynb).

Key components:

• Patient Features: Age, Sex, Blood Pressure (BP), Cholesterol levels, and Sodium-to-Potassium ratio ($\text{Na/K}$).
• Data Scaling & Preprocessing: Feature standardization (StandardScaler), One-Hot Encoding for categorical features, and train-test splitting.
• Model Optimization: Elbow method for optimal $K$-neighbor hyperparameter selection.
• Evaluation: Accuracy scores, Classification Report (Precision, Recall, F1-Score), and Confusion Matrix heatmap visualization.

---

📁 Repository Structure

Analysis-of-Drug-Types-for-Diabetes-Management/
├── KNN.ipynb     # Primary Jupyter notebook for KNN classification
├── drug200.csv   # Patient clinical features & drug label dataset
└── README.md     # Project documentation

---

🛠 Prerequisites & Installation

git clone https://github.com/amar-codes02/Analysis-of-Drug-Types-for-Diabetes-Management-Using-a-Machine-Learning-Approach.git
cd Analysis-of-Drug-Types-for-Diabetes-Management-Using-a-Machine-Learning-Approach

pip install numpy pandas matplotlib seaborn scikit-learn jupyter

---

🚀 Usage

jupyter notebook KNN.ipynb

---

📄 License

Licensed under the MIT License.