Machine-Learning

This repo contains all the projects and assignments that I have completed as a part of Machine Learning course on Coursera. The course covers pretty much all important ML concepts and algorithms and provides many excellent assignments along with quizzes, so as to strengthen one's understanding. This course discusses about many state-of-the-art Supervised and Unsupervised learning algorithms, and also does quite in-depth analysis of Recommender Systems models.

The solutions have been uploaded for personal academic purposes and can be used as a reference by others (but I strongly recommend trying out the assignments on your own).

Instructor: Prof. Andrew Ng

Offered by: Stanford University

Language: Octave (files in this repo use this) or MATLAB

Course Certificate: Verified by Coursera

Assignments

1. Linear Regression: Implementing basic linear regression model, first with single variable, and later with multiple variables.

2. Logistic Regression: Building a simple logistic regression model for binary classification.

3. Neural Networks Hand-Written Digit Recognition: Using a pre-trained 2-layer Neural Network for hand-written digit recognition.

4. BackProp Algorithm: Building and training a 2-layer NN by implementing the back-propagation algorithm to calculate the gradients.

5. Regularization: Implementing polynomial regression and using regularization to tackle overfitting due to increased number of features.

6. SVM Spam Classifier: Building a Support Vector Machine (SVM) model using Gaussian Kernel and using it to classify emails as spam vs non-spam.

7. K-means Clustering and PCA: Firstly, implementing the K-Means Clustering Algorithm and using it for image compression, and secondly, building a Principal Components Analysis (PCA) model for dimensionality reduction of a given dataset of faces for face recognition.

8. Anomaly Detection and Recommender Systems: Firstly, implementing the Anomaly Detection Algorithm for detecting anomalous examples in a given dataset of servers on a network, and secondly, using a Collaborative Filtering Algorithm to build a Recommender System for recommending movies to the users based on their reviews.