Column pdfs take a seed to give a specific sample. When calculating the fitness of an individual, that is on a particular dataset. It would be more interesting and inline with the purpose of this project to calculate fitness on the family of datasets that individual represents.
Instead, iterate over several seed and take fitness of individual from those separate fitness scores. How to do that well is another issue.
daffidwilde
commented
6 years ago
Column pdfs take a seed to give a specific sample. When calculating the fitness of an individual, that is on a particular dataset. It would be more interesting and inline with the purpose of this project to calculate fitness on the family of datasets that individual represents.
Instead, iterate over several seed and take fitness of individual from those separate fitness scores. How to do that well is another issue.