Logistic Regression: Visualization

Prepared by Sajjad Ahmed Niloy

This visualization provides an interactive, step-by-step visualization of a Logistic Regression model learning to predict diabetes. The algorithm iteratively adjusts its internal parameters to find an optimal decision boundary that separates patients into two classes. The model's predictions are based on three key health metrics from the dataset: Glucose Level, Body Mass Index (BMI), and Age. Use the controls to watch the training process unfold and see how the final model performs on unseen test data.

Controls

Learning Rate (α): 0.001

Speed: Normal

Max Epochs: 50

Training Data

Glucose	BMI	Age	Diabetes (Y)

Feature Correlation Heatmap

🧠 Model Internals

Weight 1 (Glucose): N/A

Weight 2 (BMI): N/A

Weight 3 (Age): N/A

Bias (b): N/A

Epoch: 0 / 50

Current Patient: N/A

Live Calculations

Z = (w·x) + b:...

ŷ = σ(Z):...

Loss:...

Click "Start Training" to initialize the model.

Data Distribution (Glucose vs BMI, Age as Size)

Cost Function Over Iterations