PIMA Diabetes Prediction with Deep Learning

This project uses the PIMA Indian Diabetes Dataset to build a deep learning model for predicting the onset of diabetes. The model leverages neural networks implemented using Keras and TensorFlow, and is evaluated based on several performance metrics like accuracy, AUC score, and confusion matrix.

Table of Contents

• Project Overview
• Dataset
• Dependencies
• Project Structure
• Usage
• Model Summary
• Results

Project Overview

The aim of this project is to build a binary classification model to predict whether or not a patient is likely to develop diabetes based on certain diagnostic measurements. The model uses features like glucose levels, BMI, and age to make predictions. We use a deep learning approach with multiple hidden layers and dropout regularization to prevent overfitting and enhance model performance.

Dataset

The dataset used is the PIMA Indian Diabetes Dataset from the UCI Machine Learning Repository. It consists of 768 samples with 8 features related to diagnostic measurements and one target column (Outcome).

• Pregnancies: Number of times pregnant
• Glucose: Plasma glucose concentration after 2 hours in an oral glucose tolerance test
• BloodPressure: Diastolic blood pressure (mm Hg)
• SkinThickness: Triceps skinfold thickness (mm)
• Insulin: 2-Hour serum insulin (mu U/ml)
• BMI: Body mass index (weight in kg/(height in m)^2)
• DiabetesPedigreeFunction: Diabetes pedigree function
• Age: Age in years
• Outcome: Class variable (0 if non-diabetic, 1 if diabetic)

You can download the dataset from here or use the link included in the code for fetching the data.

Dependencies

To run the code, you need the following Python libraries:

• pandas
• numpy
• tensorflow
• scikit-learn
• matplotlib
• seaborn

Install them using:

pip install pandas numpy tensorflow scikit-learn matplotlib seaborn

Project Structure

The project is structured as follows:

├── pima_diabetes_prediction.py   # Main Python script
├── README.md                     # Project documentation

Usage

To run the project, follow these steps:

1 - Clone the repository:

git clone https://github.com/your-username/pima-diabetes-prediction.git
cd pima-diabetes-prediction

2 - Run the Python script:

python pima_diabetes_prediction.py

Model Summary

The model architecture consists of the following layers:

• Input Layer: 8 input features (after scaling)
• First Hidden Layer: 128 neurons with ReLU activation
• Second Hidden Layer: 64 neurons with ReLU activation + Dropout (0.3)
• Third Hidden Layer: 32 neurons with ReLU activation + Dropout (0.3)
• Output Layer: 1 neuron with sigmoid activation for binary classification
  The model uses Adam optimizer, binary crossentropy as the loss function, and accuracy and AUC as metrics.

Results

After training the model, we evaluate it on the test set using several metrics:

• Accuracy: The model achieves an accuracy of approximately 75% on the test set.
• AUC Score: The AUC score of the model is around 0.83, indicating good performance in distinguishing between diabetic and non-diabetic patients.
• Confusion Matrix: The confusion matrix gives insight into the number of true positives, true negatives, false positives, and false negatives.

Sample Results:

• Confusion Matrix:

[[128  19]
 [ 32  52]]

• Classification Report:

              precision    recall  f1-score   support

           0       0.80      0.87      0.83       147
           1       0.73      0.62      0.67        84

    accuracy                           0.77       231
   macro avg       0.76      0.74      0.75       231
weighted avg       0.77      0.77      0.77       231