Surrogate-Assisted Cooperative Lion Swarm Optimization Algorithm for High-Dimensional Feature Selection

With the development of information science, a large amount of data has poured into people's lives. As one of the effective means of extracting massive data, feature selection has been concerned by a large number of scholars. Feature selection is an NP-hard problem, and one of the traditional methods is to use optimization algorithms to search. However, this traditional method faces serious challenges with the increase of feature- dimensional data. This is mainly because as the feature dimension becomes larger, the search space of the optimization algorithm will increase exponentially, which will seriously degrade the search performance. Previous studies have shown that evolutionary algorithms assisted by surrogate models can solve high-dimensional feature selection well. Following this research line of thought, this paper proposes a surrogate-assisted cooperative lion swarm optimization algorithm for high- dimensional feature selection, which combines co-evolution and surrogate assistance with the lion swarm optimization algorithm to improve the accuracy of high-dimensional feature selection problems. Experiments show that the algorithm can show excellent performance in feature selection problems up to 6000 dimensions.