Gradient descent and normal equations on cost function minimization for online predictive using linear regression with multiple variables

The cost function minimization is essential in finding a good model for linear regression. This paper works on prototyping and examining the minimizing cost function's two known algorithms for online predictive, namely gradient descent and normal equations. The data used in this paper are found in Open Data and split into three parts, training, test, and cross validation datasets. Empirical results are given on number of datasets, showing that normal equation performs better than gradient descent (with cross correlation 0.0117 higher and relative absolute error 0.5154 less).