Genetic algorithms are great for choosing the functions used by machine learning algorithms.
There are too many functions in the data set many times, which makes it too noisy, so we sometimes deal with this by reducing the dimension of the problem, choosing only a subset of the functions. For this purpose, genetic algorithms in many cases work (empirically), and its use is widespread both in research and in real applications with a very high dimension.
For example: text analysisis, with a bag of word models : Sometimes we want to use only a subset of words, and not all of them, so we use an algorithm for selecting features for it, and GA is one of them.
source share