MATLAB implementation of Gradient Descent algorithm for Multivariable Linear Regression. This code example includes,
- Feature scaling option
- Choice of algorithm termination based on either gradient norm tolerance or fixed number of iterations
- Randomized feature vector with randomized exponents (the exact functional relationship is not linear but with random powers of feature vectors)
- Choice of number of observations and features
- Choice of 'learning rate' and number of iterations
- Final table of actual functional coefficients and linear regression coefficients side by side