Feature selection for monotonic classification via maximizing monotonic dependency

Abstract

Monotonic classification is a special task in machine learning and pattern recognition. As to monotonic classification, it is assumed that both features and decision are ordinal and there is the monotonicity constraints between the features and decision. Little work has been focused on feature selection for this type of tasks although a number of feature selection algorithms have been introduced for nominal classification problems. However these techniques can not be applied to monotonic classification as they do not consider the monotonicity constraints. In this work, we present a technique to compute the quality of features for monotonic classification. Using gradient directing search method, this method trains a feature weight vector by maximizing the fuzzy monotonic dependency, which was defined in fuzzy preference rough sets. We conduct some experiments to compare the classification performances of the proposed method with some other techniques. The experimental results show the effectiveness of the proposed algorithm.

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

“…This is rarely the case in real-life scenarios, where class noise and discrepancies are common. Therefore, data preprocessing [21,33,8,22] and relabeling strategies [35,16] must be used to remove non-monotonic samples or to change their class labels in order to force a monotonic set.…”

Hyperrectangles Selection for Monotonic Classification by Using Evolutionary Algorithms

“…There are two steps to treat with monotonic classification problems. The first one is to preprocess the data 6 in order to "monotonize" the data set 7 , rejecting the examples that violate the monotonic restrictions or selecting features to improve classification performance and avoid overfitting 8,9 ; and the second one is to force learning only monotone classification functions. Proposals of this type are: classification trees and rule induction 10,11,12,13 , neural networks 14 and instance-based learning 15,16,17 .…”