Popular Neural Networks Classification Algorithms implemented intuitively on Matlab and Python

1. Kohonen Self Organizing Map showcasing weight updation to classify Iris Data Set into 3 categories Details: https://archive.ics.uci.edu/ml/datasets/Iris

2. Logistic Regression

3. Linear and Multivariate Polynomial Regression

4. Reinforcement Learning Solutions for the book: Reinforcement Learning: An Introduction by Richard S. Sutton and Andrew G. Barto

Logistic Regression applied mainly from Andrew Ng's CS229 lecture notes and online tutorials. Applied on IRIS data for 2 features

Details of tutorials: http://nbviewer.jupyter.org/github/tfolkman/learningwithdata/blob/master/Logistic%20Gradient%20Descent.ipynb http://aimotion.blogspot.in/2011/11/machine-learning-with-python-logistic.html

This is a sample data set provided by HackerRank for the prediction of Office Prices based on 2 or more features. I have applied Multivariate Linear Regression and Polynomial Regression on this set using Normal Equation of Least Squares method in Matrix form.

Details of the Challenge: https://www.hackerrank.com/challenges/predicting-office-space-price

Multi Armed Bandit Solutions:

Details of the Plot: Reward Values over 1000 trials averaged for:

1. Greedy (epsilon=0)
2. Epsilon-Greedy (epsilon=0.01)
3. Epsilon-Greedy (epsilon=0.1)
4. Softmax (incomplete, temperature=0.00001)