Nonparametric Monotone Classification with MOCA

Authors:
Nicola Barile;Ad Feelders
Affiliations:
-;-
Venue:
ICDM '08 Proceedings of the 2008 Eighth IEEE International Conference on Data Mining
Year:
2008

Citing 0
Cited 7

Rule learning with monotonicity constraints

ICML '09 Proceedings of the 26th Annual International Conference on Machine Learning
Isotonic Classification Trees

IDA '09 Proceedings of the 8th International Symposium on Intelligent Data Analysis: Advances in Intelligent Data Analysis VIII
Loss optimal monotone relabeling of noisy multi-criteria data sets

Information Sciences: an International Journal
Optimal restoration of stochastic monotonicity with respect to cumulative label frequency loss functions

Information Sciences: an International Journal
Citation-based journal ranks: The use of fuzzy measures

Fuzzy Sets and Systems
Monotone instance ranking with MIRA

DS'11 Proceedings of the 14th international conference on Discovery science
Large-margin feature selection for monotonic classification

Knowledge-Based Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

We describe a monotone classification algorithm called MOCA that attemptsto minimize the mean absolute prediction error for classification problems with ordered class labels.We first find a monotone classifier with minimum L1 loss on the training sample, and then use a simpleinterpolation scheme to predict the class labels for attribute vectors not present in the training data.We compare MOCA to the Ordinal Stochastic Dominance Learner (OSDL), on artificial as well asreal data sets. We show that MOCA often outperforms OSDL with respect to mean absolute prediction error.