Dynamicity Analysis in the Selection of Classifier Ensembles Parameters

• Jesaías Carvalho Pereira Silva UFRN
• Anne Magaly de Paula Canuto UFRN
• Araken de Medeiros Santos UFERSA

Resumo

Over the years, significant progress has been made in the realm of classifier ensembles research. Several methods to enhance their efficiency have been proposed, applicable to both homogeneous and heterogeneous ensemble structures. A key challenge in employing classifier ensembles lies in determining their structure (hyper-parameters). Basically, the ensemble structure selection can be done in two different ways, static and dynamic selection. Unlike static selection, dynamic selection defines the ensemble structure for each testing instance. Different dynamic selection methods have been proposed in the literature, mainly for ensemble members and features, but very little effort has been done to propose dynamic selection methods for combination methods. Therefore, it is important to evaluate the impact of a dynamic selection of combination methods or both (methods and members) in the creation of robust classifier ensembles. In this paper, an exploratory analysis of dynamic selection of the main ensemble structure parameters will be performed. In order to do this, three different scenarios will be assessed: Full Static ensemble, Partial Dynamic ensemble and Full Dynamic ensemble. Finally, an empirical analysis of these scenarios will be carried out. Our findings show that the use of a full dynamic selection provides more robust classifier ensembles, in most of the analyzed cases.