Research on Feature Selection Algorithm Based on Relief-F and Genetic Algorithm for Underwater Target

Abstract

Aiming at the problem of multi-domain feature selection for underwater target, the feature selection algorithm based on Relief-F and genetic algorithm is proposed. Relief-F algorithm, reduce the search space and speed up the search speed of genetic algorithm. The population initialization of the genetic algorithm is guided by the result of feature sorting. Classification experiments show that the algorithm have a better comprehensive performance, which can get a higher recognition rate with fewer features.