Regresar

Variation of the Intercession Coefficient Used as a Hyper Parameter in Machine Learning in Regression Models

Abstract:

Within the area of data science, the hyper parameters arguments affect the execution of the algorithms and due to their particularity, must be used separately for each machine learning model in quantitative pbkp_redictions, known as regression. Two quality metrics from regression models will be used in order to demonstrate the changes: Mean Square Error, (MSE) and the value of the R2 coefficient. In the present work, using simulation of the interception coefficient, it will be demonstrated the existence of machine learning algorithms that are sensitive: Mean Square Error and/or the value of the R2 coefficient. It is important to highlight that the intercept coefficient is considered as a reference argument. The present research is only a very small space of an automated machine learning process (AutoML) with regard to sensitivity analysis.