Machine Learning !!!
This repository is about solving ML problems using different algorithms. It has a detailed overview of algorithms used and their implementation from scratch using Python's Numpy library.
But Before using or even learning machine learning, the top 2 prerequisites to be mastered are Linear Algebra and Probability theory. And it imperative to preprocess the data and draw useful patterns from data.
Data Visualization notebook provides a review of linear albegra and probability theory. It illustrates the process of data processing and drawing preliminary observations from two datasets- one for the regression problem and another for the classification problem. The attributes of these datasets are visualized using bar plots, box plots, boxen plots, Facetgrid, joint plot, heatmap, pairgrid, count plot and different flavors of the scatterplot. Seaborn and Matplotlib libraries are used.
Regression Algorithm notebook provides a review of regression algorithms - Least Squares and Least Mean Squares. Both of these algorithms are implemented. And a regression problem "Predicting the traffic" is solved using both of these algorithms and their performance metrics are analyzed.
Classification Algorithm notebook provides a review of classification algorithms - Pocket Algorithm, Quadratic Discriminant Analysis, Linear Discriminant Analysis, and Logistic Regression. All of these are then implemented and the classification problem of "whether an online user would make a purchase" is solved using them. This notebook has review classification metrics. And the resultant metrics for different algorithms are analyzed.