Diabetes Detection App

This project is a Python-based GUI application to detect diabetes using a custom logistic regression model. The application reads healthcare data, trains a machine learning model, and allows users to input patient data to predict the likelihood of diabetes.

Features

• Custom Logistic Regression Model: Implements logistic regression from scratch.
• GUI Interface: User-friendly interface created with PyQt5 and Qt Designer.
• Data Normalization: Scales input data using MinMaxScaler to ensure accurate predictions.
• Live Predictions: Users can input data to get immediate predictions about diabetes.

Dataset

The application uses a dataset named Healthcare-Diabetes.csv, which should have the following columns:

1. Pregnancies
2. Glucose
3. BloodPressure
4. SkinThickness
5. Insulin
6. BMI
7. DiabetesPedigreeFunction
8. Age
9. Outcome (binary: 0==> salekt'ha for no diabetes, 1==> rak 3andk el sokkker for diabetes)

Make sure the dataset is in the same directory as the application.

Installation

1. Clone the repository:

   git clone https://github.com/your-repo/diabetes-detection-app.git

2. Install dependencies:

   python pip install -m PyQt5 pandas numpy scikit-learn

How to run

1. Open a terminal in the project directory
2. Run the application:

   python app.py

3. The GUI window will open. Input patient data and click the "Predict" button to detect diabetes.

File Descriptions

• app.py: Main application file containing the GUI and logic for the diabetes detection model.
• diabetes.ui: The Qt Designer file defining the graphical interface.
• Healthcare-Diabetes.csv: Dataset used for training and testing the model.
• README.md:Documentation for the project.

How it works

1. Data Preprocessing
   • The dataset is split into training (80%) and testing (20%) sets.
   • he features are normalized using MinMaxScaler for better model performance.
2. Model Training
   • A logistic regression model is implemented and trained from scratch.
   • The training process minimizes the binary cross-entropy loss.
3. Prediction
   • Input features are normalized using the same scaler as the training data.
   • The logistic regression model predicts the probability of diabetes.
4. GUI
   • The user inputs values for features like glucose levels, blood pressure, and BMI.
   • The application displays whether the patient is likely to have diabetes.

EXAMPLE

1. Input

• Pregnancies: 5
• Glucose: 140
• Blood Pressure: 80
• Skin Thickness: 20
• Insulin: 100
• BMI: 28.5
• Diabetes Pedigree Function: 0.5
• Age: 45

2. Output

• rak 3andk el sokkker (likely diabetic)
• or salekt'ha (not diabetic)

Auther

Hakim Saoud [email protected] https://www.linkedin.com/in/hakim-saoud/