🩺 Diabetes Prediction using K-Nearest Neighbors (KNN)

📌 Project Overview

This project aims to predict whether a patient has diabetes using the Pima Indians Diabetes Dataset and the K-Nearest Neighbors (KNN) classification algorithm.

The project covers the complete machine learning workflow, from data preprocessing to model evaluation.

---

📂 Dataset

• Dataset: Pima Indians Diabetes Dataset
• Source: Kaggle / UCI Machine Learning Repository

Features

• Pregnancies
• Glucose
• BloodPressure
• SkinThickness
• Insulin
• BMI
• DiabetesPedigreeFunction
• Age

Target

• 0: Non-Diabetic
• 1: Diabetic

---

🛠 Technologies Used

• Python
• Pandas
• NumPy
• Matplotlib
• Seaborn
• Scikit-learn

---

📋 Machine Learning Workflow

1. Data Loading
2. Data Exploration
3. Data Cleaning
4. Train/Test Split
5. Feature Scaling using StandardScaler
6. KNN Model Training
7. Hyperparameter Tuning (Best K Selection)
8. Model Evaluation
   • Accuracy
   • Confusion Matrix
   • Classification Report

---

📊 Results

Best K value:

K = 17

Model Performance:

Metric	Score
Accuracy	77%
Precision (Diabetic)	72%
Recall (Diabetic)	56%
F1-Score (Diabetic)	63%

---

📚 What I Learned

During this project I learned how to:

• Build a classification model using KNN.
• Split data into training and testing sets.
• Standardize numerical features.
• Tune hyperparameters.
• Evaluate a classification model using:
  • Accuracy
  • Precision
  • Recall
  • F1-score
  • Confusion Matrix

---

🚀 Future Improvements

• Try Decision Tree Classifier.
• Compare multiple classification algorithms.
• Improve recall for diabetic patients.
• Deploy the model using Streamlit.

---

👨‍💻 Author

Taha Lazar