Regression is simple and efficient statistics learning algorithm, used as a baseline for other more complex methods. This repository contains the code and algorithm of linear regression, regularization and logistic regression for binary and multiclass classification.
Logistic regression implements the sigmoid function for binary classification and softmax for multiclass classification. Gradient descent is used to calculate the variable learning rate.
code.py- Python code.
Report - The report describes the theoretical fundamentals, datasets, algorithm and result analysis. Contains the PDF and R Markdown files.
diabetes: Dataset used to train the model.