ML Models From Scratch

This repository contains implementations of fundamental machine learning algorithms from scratch using Python and NumPy. It's designed for learning purposes and interview preparation.

Inspiration

Implemented Algorithms

• ☑️ K-Means Clustering
• ☑️ Linear Regression
• ☑️ Logistic Regression
• ☑️ Decision Trees
• SVD
• PCA
• Support Vector Machine
• Neural Networks

Linear Regression Evaluation

Dataset: Single Feature (Attendance Hours → Final Marks)

Model	MSE	R² Score
Custom Linear Regression	0.1725	0.8234
Sklearn SGDRegressor	0.1725	0.8233

Dataset: Multiple Features — Graduate Admission Prediction

Model	MSE	R² Score
Custom Linear Regression	0.5535	0.5652
Sklearn SGDRegressor	0.2349	0.8155

Logistic Regression Evaluation (User Metadata -> Purchase Prediction)

Model	MSE	R² Score
Custom Logistic Regression	3.13 × 10⁻³⁰	1.0
Sklearn Logistic Regression	0.0	1.0