Fuzzy k-Nearest Neighbors with monotonicity constraints: Moving towards the robustness of monotonic noise

This paper proposes a new model based on Fuzzy k-Nearest Neighbors for\nclassification with monotonic constraints, Monotonic Fuzzy k-NN (MonFkNN).\nReal-life data-sets often do not comply with monotonic constraints due to class\nnoise. MonFkNN incorporates a new calculation of fuzzy memberships, which\nincreases robustness against monotonic noise without the need for relabeling.\nOur proposal has been designed to be adaptable to the different needs of the\nproblem being tackled. In several experimental studies, we show significant\nimprovements in accuracy while matching the best degree of monotonicity\nobtained by comparable methods. We also show that MonFkNN empirically achieves\nimproved performance compared with Monotonic k-NN in the presence of large\namounts of class noise.\n