Multiobjective Firefly Algorithm for Variable Selection in Multivariate Calibration

摘要

Firefly Algorithm is a newly proposed method with potential application on several real world problems, such as variable selection problem. This paper presents a Multiobjective Firefly Algorithm (MOFA) for variable selection in multivariate calibration models. The main objective is to propose an optimization to reduce the error value prediction of the property of interest, as well as reducing the number of variables selected. Based on the results obtained, it is possible to demonstrate that our proposal may be a viable alternative in order to deal with conflicting objective-functions. Additionally, we compare MOFA with traditional algorithms for variable selection and show that it is a more relevant contribution for the variable selection problem.