@inproceedings{ceb3f68f8fb345f3a1188f971a63c2ac,
title = "Detection of CAN by ensemble classifiers based on ripple down rules",
abstract = "It is well known that classification models produced by the Ripple Down Rules are easier to maintain and update. They are compact and can provide an explanation of their reasoning making them easy to understand for medical practitioners. This article is devoted to an empirical investigation and comparison of several ensemble methods based on Ripple Down Rules in a novel application for the detection of cardiovascular autonomic neuropathy (CAN) from an extensive data set collected by the Diabetes Complications Screening Research Initiative at Charles Sturt University. Our experiments included essential ensemble methods, several more recent state-of-the-art techniques, and a novel consensus function based on graph partitioning. The results show that our novel application of Ripple Down Rules in ensemble classifiers for the detection of CAN achieved better performance parameters compared with the outcomes obtained previously in the literature.",
author = "Andrei Kelarev and Richard Dazeley and Andrew Stranieri and John Yearwood and Herbert Jelinek",
note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2012.; 12th International Workshop on Knowledge Management and Acquisition for Intelligent Systems, PKAW 2012, held in conjunction with 12th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2012 ; Conference date: 05-09-2012 Through 06-09-2012",
year = "2012",
doi = "10.1007/978-3-642-32541-0_13",
language = "British English",
isbn = "9783642325403",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "147--159",
editor = "Deborah Richards and Kang, {Byeong Ho}",
booktitle = "Knowledge Management and Acquisition for Intelligent Systems - 12th Pacific Rim Knowledge Acquisition Workshop, PKAW 2012, Proceedings",
address = "Germany",
}