Implementation of various machine learning models and their applications from scratch

The package 'ml' includes K-Nearest Neighbors, Linear regression, Logistic regression, Neural Network

K-Nearest Neighbor decision boundary visualization

knn_vis.ipynb Visualization of decision boundaries for classification of a 2-D dataset.

Breast Cancer Wisconsin Data Set

breast_cancer_wisconsin.ipynb Predict whether a cancer is benign or malignant. A logistic regression model is trained using the dataset available on https://www.kaggle.com/uciml/breast-cancer-wisconsin-data. I am able to get about 95% accuracy on an untrained dataset.

Handwritten digit classification using the MNIST dataset

MNIST.ipynb A multiclass classifier to recognize handwritten digits using the MNIST dataset available on https://www.kaggle.com/oddrationale/mnist-in-csv. I am able to get about 91% accuracy on an untrained dataset using a simple feed-forward neural network