view release on metacpan or  search on metacpan

docs/specifications.t  view on Meta::CPAN

#   ? Data processing: splitting data, k-fold
#   -...
#
#
############ "flow" of the codes ############

# these three steps could be done in seperated scripts if necessary
# &train and &validate could be put inside a loop or something
# the parameters make more sense when they are taken from @ARGV
    # so when it's the first time training, it will create the nerve_file,
    # the second time and up it will directly overrride that file since everything is read from it
    # ... anyway :) afterall training stage wasn't meant to be a fully working program, so it shouldnt be a problem
# just assume that 
$perceptron->train( $stimuli_train, $save_nerve_to_file ); 
    # reads training stimuli from csv
    # tune attributes based on csv data
        # calls the same subroutine to do the calculation
    # shouldn't give any output upon completion
    # should save a copy of itselt into a new file
    # returns the nerve's data filename to be used in validate()
        # these two can go into a loop with conditions checking