Abstract: Feature or variable selection still remains an unsolved problem, due to the infeasible evaluation of all the solution space. Several algorithms based on heuristics have been proposed so far with successful results. However, these algorithms were not designed for considering very large datasets, making their execution impossible, due to the memory and time limitations. This paper presents an implementation of a genetic algorithm that has been parallelized using the classical island approach, but also considering graphic processing units to speed up the computation of the fitness function. Special attention has been paid to the population evaluation, as well as to the migration operator in the parallel genetic algorithm (GA), which is not usually considered too significant; although, as the experiments will show, it is crucial in order to obtain robust results.
Keywords: general-purpose computing on graphics processing units (GPGPU); feature selection; variable selection; big data
Export to BibTeX
MDPI and ACS Style
Guillén, A.; García Arenas, M.I.; van Heeswijk, M.; Sovilj, D.; Lendasse, A.; Herrera, L.J.; Pomares, H.; Rojas, I. Fast Feature Selection in a GPU Cluster Using the Delta Test. Entropy 2014, 16, 854-869.
Guillén A, García Arenas MI, van Heeswijk M, Sovilj D, Lendasse A, Herrera LJ, Pomares H, Rojas I. Fast Feature Selection in a GPU Cluster Using the Delta Test. Entropy. 2014; 16(2):854-869.
Guillén, Alberto; García Arenas, M. I.; van Heeswijk, Mark; Sovilj, Dusan; Lendasse, Amaury; Herrera, Luis J.; Pomares, Héctor; Rojas, Ignacio. 2014. "Fast Feature Selection in a GPU Cluster Using the Delta Test." Entropy 16, no. 2: 854-869.