bhavyejain / ml-coursera-matlab Goto Github PK

This repository contains my solutions to the coding assignments of the Coursera Machine Learning course (Stanford University).

Linear-regression folder contains an implementation of the Linear Regression algorithm. (Week 2) The algorithm has been implemened by using Gradient Descent as well as the Normal Equation. Feature scaling was also made use of.

Logistic-regression folder contains an implementation of the Logistic Regression algorithm. (Week 3) The algorithm has been implemented in both, a regularized as well as an unregularized fasion.

Multi-class-classification-and-neural-networks folder (week 4) contains my implementation of one-vs-all method of multi class classification. As the second part of the exercise, there a code for the feedforward propagation algorithm for an already trained neural network for handwritten digit recognition.

Neural-network-learning folder (week 5) contains the calcolation of cost by forward propagation and implementation of backpropagation algorithm to compute the partial derivatives for a neural network. Finally, The gradients are used to optimise the weights for the neural network.

Bias-and-variance folder (week 6) contains the implementation of regularized linear regression, regularized polynomial regression, and learning curves for the algorithm.

Support-vector-machines (week 7) contains an implementation of SVM's with the linear and gaussian kernels. In the second part, I there is an implementation of an e-mail spam classifier using a linear kernal SVM.

K-means-and-pca (Week 8) contains my implemetation of the K-means clustering algorithm and principal component analysis. This marked the beginning of unsupervised learning in the course. We used PCA to compress and reconstruct an image.

Anomaly-detections-and-recommender-systems (Week 9) concludes the programming assignments of the course. Here I used Gaussian Probability distribution for detecting anomalies in a system. In the second part, I used collaborative filtering to create a recommender system for movies.