Discussing the article: "Elastic net regression using coordinate descent in MQL5"

 

Check out the new article: Elastic net regression using coordinate descent in MQL5.

In this article we explore the practical implementation of elastic net regression to minimize overfitting and at the same time automatically separate useful predictors from those that have little prognostic power.

Cartesian


Coordinate descent is a method of optimization well suited to multivariable optimization. A complex multidimensional optimization problem is reduced into a collection of one dimensional problems.  Achieved by iteratively minimizing each of the individual dimensions of a function , whilst keeping the values of the  function in other dimensions static. There are numerous resources  on the internet that can provide more detailed explanations for  those interested. Here, we are interested in its application to  strategy development.
 
For our purposes, the coordinate descent method will be used in two  ways in the implementation of elastic net regression. First it will  be used to determine the optimal lambda based on a fixed alpha,  specified by the user. Once this is done, the optimization method is  called again to work on the beta coefficients of the regression equation. Lets dive into some code to see how this is accomplished.

Author: Francis Dube