ITMO_FS.filters.multivariate.MIMAGA

class ITMO_FS.filters.multivariate.MIMAGA(mim_size, pop_size, max_iter, f_target, k1, k2, k3, k4)

__init__(mim_size, pop_size, max_iter, f_target, k1, k2, k3, k4)

Parameters:

• mim_size – desirable number of filtered features after MIM
• pop_size – initial population size
• max_iter – maximum number of iterations in algorithm
• f_target – desirable fitness value
• k1 – consts to determine crossover probability
• k2 – consts to determine crossover probability
• k3 – consts to determine mutation probability
• k4 – consts to determine mutation probability

mimaga_filter(genes, classes)

The main function to run algorithm :param genes: initial dataset in format: features are rows, samples are columns :param classes: distribution pf initial dataset :return: filtered with MIMAGA dataset, fitness value