Topics

#accuracy-score #artificial-intelligence #classification #data-collection #data-standardization #dataset #healthcare #jupyter-notebook #learning-materials #machine-learning-algorithms #model-evaluation #model-training #numpy #pandas #prediction-system #python #scikit-learn #supervised-learning #support-vector-machine #train-test-split

Language breakdown

No language data available.

▸ Files click to expand

README

🩺 SVM Diabetes Prediction

🤖 Machine Learning Model for Diabetes Prediction using Support Vector Machine

Predicting diabetes risk using medical indicators with high accuracy SVM classification

🎯 Project Overview

This project implements a Support Vector Machine (SVM) classifier to predict diabetes in patients based on medical diagnostic measurements. The model analyzes various health indicators to determine the likelihood of diabetes, providing a valuable tool for early diagnosis and preventive healthcare.

🎯 Objectives

Build an accurate diabetes prediction model using SVM
Analyze medical features that contribute to diabetes risk
Provide a reliable tool for healthcare screening
Demonstrate machine learning applications in medical diagnosis

🏥 Medical Significance

Early Detection: Helps identify diabetes risk before symptoms appear
Preventive Care: Enables timely intervention and lifestyle modifications
Healthcare Efficiency: Assists medical professionals in screening processes
Data-Driven Decisions: Provides objective risk assessment based on measurable factors

🔬 Dataset Information

The project uses the Pima Indian Diabetes Database, a well-known medical dataset for diabetes prediction research.

📊 Dataset Statistics

Total Samples: 768 patients
Features: 8 medical predictor variables
Target Classes: Binary (Diabetic/Non-Diabetic)
Class Distribution:

- Non-Diabetic: 500 instances (65.1%) - Diabetic: 268 instances (34.9%)

🩺 Medical Features

| Feature | Description | Unit | Range | | ---------------------------- | ---------------------------- | ------- | ---------- | | Pregnancies | Number of times pregnant | Count | 0-17 | | Glucose | Plasma glucose concentration | mg/dL | 0-199 | | BloodPressure | Diastolic blood pressure | mmHg | 0-122 | | SkinThickness | Triceps skin fold thickness | mm | 0-99 | | Insulin | 2-Hour serum insulin | mu U/ml | 0-846 | | BMI | Body mass index | kg/m² | 0-67.1 | | DiabetesPedigreeFunction | Diabetes pedigree function | Score | 0.078-2.42 | | Age | Age of patient | Years | 21-81 |

🎯 Target Variable

Outcome:

- 0 = Non-Diabetic - 1 = Diabetic

🚀 Features

✨ Core Functionality

🔬 Data Preprocessing: Comprehensive data cleaning and standardization
🤖 SVM Classification: Linear kernel SVM implementation
📊 Model Evaluation: Accuracy assessment on training and testing sets
🔮 Prediction System: Real-time diabetes risk prediction
📈 Performance Metrics: Detailed model performance analysis

🛠️ Technical Features

Data Standardization: StandardScaler for feature normalization
Train-Test Split: Stratified sampling for balanced evaluation
Cross-Validation Ready: Extensible for k-fold validation
Modular Design: Clean, reusable code structure

⚙️ Installation

📋 Prerequisites

Python 3.7 or higher
pip package manager
Jupyter Notebook (recommended)

🔧 Setup Instructions

Clone the Repository

git clone https://github.com/NhanPhamThanh-IT/SVM-Diabetes-Prediction.git
   cd SVM-Diabetes-Prediction

Create Virtual Environment (Recommended)

python -m venv diabetes_env # Windows diabetes_env\Scripts\activate

# macOS/Linux source diabetes_env/bin/activate

Install Required Packages

pip install numpy pandas scikit-learn jupyter matplotlib seaborn

Or using requirements.txt:

pip install -r requirements.txt

Launch Jupyter Notebook

jupyter notebook SVM-Diabetes-Prediction.ipynb

📦 Required Dependencies

numpy>=1.21.0
pandas>=1.3.0
scikit-learn>=1.0.0
jupyter>=1.0.0
matplotlib>=3.4.0  # For visualization (optional)
seaborn>=0.11.0    # For advanced plots (optional)

💻 Usage

🚀 Quick Start

Open the Jupyter Notebook

jupyter notebook SVM-Diabetes-Prediction.ipynb

Run All Cells

- Execute cells sequentially to train the model - Observe data analysis and preprocessing steps - View model training and evaluation results

🔮 Making Predictions

The model accepts 8 medical features to predict diabetes risk:

# Example prediction
import numpy as np
Input format: [Pregnancies, Glucose, BloodPressure, SkinThickness,
               Insulin, BMI, DiabetesPedigreeFunction, Age]
sample_data = np.array([[5, 166, 72, 19, 175, 25.8, 0.587, 51]])
Standardize and predict
stddata = scaler.transform(sampledata)
prediction = classifier.predict(std_data)
if prediction[0] == 0:     print("🟢 The person is not diabetic") else:     print("🔴 The person is diabetic")

📊 Model Training Process

Data Loading: Import diabetes dataset
Exploratory Analysis: Statistical summary and data distribution
Data Preprocessing: Feature standardization using StandardScaler
Data Splitting: 80-20 train-test split with stratification
Model Training: Linear SVM classifier training
Evaluation: Accuracy calculation on both training and testing sets
Prediction: Individual risk assessment system

🧠 Machine Learning Approach

🔬 Support Vector Machine (SVM)

Why SVM for Diabetes Prediction?

Linear Separability: Finds optimal decision boundary between classes
High Dimensional Data: Effective with multiple medical features
Robust to Overfitting: Generalizes well on medical datasets
Margin Maximization: Creates reliable classification boundaries

⚙️ Model Configuration

# SVM Classifier Setup
classifier = svm.SVC(kernel='linear')
Key Parameters:
- kernel='linear': Linear decision boundary
- C=1.0 (default): Regularization parameter
- gamma='scale': Kernel coefficient

🔄 Data Preprocessing Pipeline

Feature Extraction: Separate features (X) and target (Y)
Standardization: StandardScaler normalization

scaler = StandardScaler()
   Xscaled = scaler.fittransform(X)

Train-Test Split: Stratified 80-20 split

Xtrain, Xtest, Ytrain, Ytest = traintestsplit(
       X, Y, testsize=0.2, stratify=Y, randomstate=2
   )

📊 Model Performance

🎯 Accuracy Metrics

The model achieves high accuracy on both training and testing datasets:

Training Accuracy: ~78-80%
Testing Accuracy: ~75-77%
Balanced Performance: Good generalization without overfitting

📈 Performance Analysis

Strengths:

✅ High accuracy on medical diagnostic task
✅ Balanced performance across classes
✅ Fast prediction capability
✅ Interpretable linear decision boundary

Considerations:

📊 Dataset size limitations (768 samples)
🔄 Potential for ensemble improvements
📈 Room for hyperparameter optimization

📁 Project Structure

SVM-Diabetes-Prediction/
│
├── 📓 SVM-Diabetes-Prediction.ipynb   # Main Jupyter notebook
├── 📊 diabetes_data.csv               # Dataset file
├── 📝 README.md                       # Project documentation
├── 📜 LICENSE                         # MIT License
│
├── 📁 docs/                           # Documentation
│   ├── 📋 dataset.md                  # Dataset documentation
│   └── 🧠 svm-algorithm.md            # SVM algorithm details
│
└── 📁 models/                         # Saved models (optional)
    └── 🤖 svmdiabetesmodel.pkl      # Trained model

🔍 Technical Implementation

🧩 Code Architecture

1. Data Import & Analysis

# Load and explore dataset
diabetesdataset = pd.readcsv('diabetes_data.csv')
diabetes_dataset.describe()
diabetesdataset['Outcome'].valuecounts()

2. Feature Engineering

# Separate features and target
X = diabetes_dataset.drop(columns='Outcome', axis=1)
Y = diabetes_dataset['Outcome']

3. Data Standardization

# Normalize features
scaler = StandardScaler()
Xstandardized = scaler.fittransform(X)

4. Model Training

# Train SVM classifier
classifier = svm.SVC(kernel='linear')
classifier.fit(Xtrain, Ytrain)

5. Model Evaluation

# Calculate accuracy
trainaccuracy = accuracyscore(Ytrain, classifier.predict(Xtrain))
testaccuracy = accuracyscore(Ytest, classifier.predict(Xtest))

🔧 Extensibility

The codebase is designed for easy extension:

Additional Kernels: RBF, polynomial, sigmoid
Hyperparameter Tuning: Grid search implementation
Cross-Validation: K-fold validation
Feature Selection: Recursive feature elimination
Ensemble Methods: Voting classifiers

📈 Results & Analysis

📊 Key Findings

Model Effectiveness: SVM demonstrates strong performance for diabetes prediction
Feature Importance: Glucose level and BMI are strong predictors
Generalization: Model maintains consistent accuracy across train/test sets
Clinical Relevance: Results align with medical knowledge of diabetes risk factors

🎯 Medical Insights

Glucose Levels: Primary indicator of diabetes risk
BMI Correlation: Strong relationship with diabetes occurrence
Age Factor: Increasing risk with age
Family History: Diabetes pedigree function significance

🛠️ Technologies Used

🐍 Core Technologies

Python: Primary programming language
Jupyter Notebook: Interactive development environment
NumPy: Numerical computations and array operations
Pandas: Data manipulation and analysis
Scikit-learn: Machine learning algorithms and tools

📚 Libraries & Frameworks

import numpy as np              # Numerical operations
import pandas as pd             # Data manipulation
from sklearn.preprocessing import StandardScaler    # Feature scaling
from sklearn.modelselection import traintest_split # Data splitting
from sklearn import svm         # Support Vector Machine
from sklearn.metrics import accuracy_score          # Model evaluation

🔧 Development Tools

Git: Version control
Markdown: Documentation
CSV: Data storage format

🤝 Contributing

We welcome contributions to improve the diabetes prediction model! Here's how you can contribute:

🌟 Ways to Contribute

🐛 Bug Reports: Report issues or bugs
💡 Feature Requests: Suggest new features or improvements
📝 Documentation: Improve project documentation
🔬 Model Enhancement: Optimize algorithms or add new models
📊 Data Analysis: Enhance data preprocessing or visualization

📋 Contribution Guidelines

Fork the Repository
Create Feature Branch

git checkout -b feature/your-feature-name

Make Changes
Add Tests (if applicable)
Commit Changes

git commit -m "Add: Your descriptive commit message"

Push to Branch

git push origin feature/your-feature-name

Create Pull Request

🎯 Areas for Improvement

Model Optimization: Hyperparameter tuning, cross-validation
Additional Algorithms: Random Forest, Gradient Boosting, Neural Networks
Data Visualization: Enhanced plots and statistical analysis
Web Interface: Flask/Django web application
Mobile App: Mobile diabetes risk calculator
API Development: REST API for model predictions

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

MIT License Copyright (c) 2025 NhanPhamThanh-IT Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.