Gradient Descent From Scratch- Anusha Kodavuru

Gradient Descent From Scratch

A curated collection of Kaggle notebooks covering Gradient Descent from first principles to practical implementations in Machine Learning.

Learn Gradient Descent Understand the mathematics behind gradient descent by deriving parameter update rules and implementing the algorithm from scratch on synthetic regression data.
Simple Linear Regression Build a custom Gradient Descent Regressor for single-variable linear regression and learn how model parameters converge during training.
Changing Intercept Implement gradient descent by simultaneously updating both the slope and intercept, with a custom regressor built entirely from scratch.
Stochastic GD Implement Stochastic Gradient Descent by updating model parameters after every training example and compare its learning behaviour with Batch GD.
Batch GD Extend gradient descent to multivariate linear regression by training a custom model on the Diabetes dataset using vectorized parameter updates.
Mini-Batch GD Implement Mini-Batch Gradient Descent and explore how batching offers a practical balance between Batch GD and SGD.
Perceptron Loss Learn how activation functions and loss functions define different models, then implement Perceptron learning with Gradient Descent on a binary classification dataset.