Regression estimation with support vector machines : theory and application

Abstract EN

The goal of regression estimation is to construct a model that describes the relationships between a set of input and output variables.

Applications of regression estimation exist in almost every field.

This paper describes the application of a new approach, support vector machine (SVM), for constructing regression models.

This technique is based on the concepts of machine learning with the ability to handle complex problems and represent non-linear relationships.

SVM regression is used to predict the demand for construction housing in Saudia Arabia using six different variables as inputs to the model.

The performance of the model was evaluated using three different criteria which are the coefficient of determination (R2), mean squared error (MSEJ) and mean absolute error {MAE).

In addition, a comparison was made between the performance of the SVM regression and the classical multiple regression (MR).

The SVM regression model showed superior performance in modeling the relationships between variables.