2025
Neural network foundations, from the perceptron to a tuned DNN
A perceptron and gradient descent written from scratch in NumPy, and a deep regression network tuned with Keras Tuner.
Problem
Understand how a neural network learns before relying on frameworks, then apply it to a real regression problem.
Data
AND/OR truth tables for the perceptron, an analytic cost function for gradient descent, and a course-provided regression dataset (8 numeric features `x1`…`x8`, target `y`) for the DNN.
Method
A step-function perceptron for AND/OR gates, gradient descent on a cost function, and a 3-hidden-layer DNN with EDA, Keras Tuner, early stopping and 5-fold cross-validation.
Result
The regression DNN reached a test RMSE of 89.8 (5-fold CV, 92.6 ± 40.7), with fold-to-fold variability documented as a model limitation.
Stack
- Python
- NumPy
- TensorFlow
- Keras Tuner
- scikit-learn
Context
Course: Machine Learning II · Machine Learning III · 2025