Abstract:
|
The surge of interest in personalized medicine during recent years has increased the application of ordinal classification problems in biomedical science. Currently, accuracy, Kendall's tau-b, and AMAE are three commonly used metrics for evaluating the effectiveness of a multiclass ordinal classifier. Although there are benefits to each, no single metric considers the benefits of predictive accuracy with the tradeoffs of misclassification error. In addition, decisions that consider pairwise analysis of the metrics are not trivial due to inconsistent findings. Hence, a new cost-sensitive metric is proposed to find the optimal tradeoff between the two most critical performance measures of a classification task. The proposed method accounts for imbalanced class distribution, the inherent ordinal data structure, and misclassification cost, thereby providing a more comprehensive tool for comparative analysis of multiple ordinal classifiers. The strengths of our methodology are demonstrated through real data analysis of cancer datasets and simulation studies.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.