A chaotic parallel antlion optimization algorithm for feature selection
Xun Zhou, Hongwei Chen, Dewei Shi · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2022
Feature selection is an important part of data preprocessing. There are currently three types of feature selection, namely filter, wrapper, and embedded feature selection methods. The latter is a combination of the former two. However, when the feature dimension is high and the amount of data is large, the filter method may not be able to extract a better subset, ignoring the interaction between the features, and the embedded algorithm has requirements for the learning model. Currently, wrapper algorithms face the problem of high computational overhead. The Spark platform based on in-memory computing has advantages in handling iterative tasks. Therefore, this paper proposes a chaotic parallel antlion optimization algorithm for feature selection (QPSALO) based on Spark, and conducts experimental comparisons on seven datasets. The results show that the QPSALO algorithm can improve the quality of candidate feature subsets while speeding up the execution.