Diabetes Prediction Using SVM

Problem Statement

The objective of this project is to build a machine learning model that predicts whether a patient has diabetes based on medical diagnostic measurements.

Tech Stack

• Python
• Pandas, NumPy
• Scikit-learn
• Support Vector Machine (SVM)

Approach

• Loaded and analyzed the diabetes dataset
• Split the dataset into training and testing sets
• Applied feature scaling using StandardScaler
• Trained a linear Support Vector Machine classifier
• Evaluated the model using accuracy metric
• Ensured no data leakage by fitting the scaler only on training data

Project Structure

• dataset/ – Contains the diabetes dataset
• src/
  • data_preprocessing.py – Data loading, splitting, and scaling
  • model.py – SVM model definition
  • train.py – Model training and evaluation
• requirements.txt – Project dependencies
• README.md – Project documentation

Results

• Achieved an accuracy of 77.2% on the test dataset

How to Run

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt

3. Run the training script:

   python src/train.py