Use of very small training data subsets in parallel distributed genetic fuzzy rule selection | IEEE Conference Publication | IEEE Xplore