Cost curves: An improved method for visualizing classifier performance

By C. Drummond, Robert Craig Holte. Machine-mediated learning, 2006

364 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/71708ce984e0896e7383435913547e770572410e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

Citations in the corpus

Rule Extraction from Support Vector Machines: A Sequential Covering Approach (2007, background): “For example, a number of recent studies [15], [16], [17] have demonstrated the efficacy of alternative methods such as cost curves [15], precision-recall operating characteristic (P-ROC) curves [16], and techniques that measure failure avoidance or class discrimination [17], which are clearly…”
Training and assessing classification rules with imbalanced data (2012, methodology): “Similarly, precision-recall curves may be adopted for assessing the classification accuracy (Davis and Goadrich, 2006) and cost curves feature the ability to compare the performance of a classifier over a range of misclassification costs and class distributions (Drummond and Holte, 2006). For a comp”
An insight into classification with imbalanced data: Empirical results and current trends on using data intrinsic characteristics (2013, methodology): “Furthermore, the expected cost or profit of each model might be analyzed using cost curves [40], lift and ROI graphs [83].”
A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models. (2019, neutral): “For example cost curves are very similar to decision curve analysis [102].”
A Review On Evaluation Metrics For Data Classification Evaluations (2015, background): “Although these metrics are better than accuracy or error rate, its graphical-based output limits a metric such as Receiver Operating Curve (ROC) [4], Bayesian Receiver Operating Characteristic (B-ROC) [1], Precision-Recall Curve [5], cost curve [6], lift and chart calibration plot [35] to be employe”
A study of statistical techniques and performance measures for genetics-based machine learning: accuracy and interpretability (2009, background): “On the other hand, the performance of classifiers is not only given by their classification rate and there is a growing interest in proposing or adapting new accuracy measures (Ben-David 2007; Drummond and Holte 2006).”
Interactive optimization for steering machine classification (2010, methodology): “These include the use of summary statistics of classifiers such as log scores, classification accuracies, F-measures, and a variety of other statistics [24,26], ROC curves, area under ROC, and cost curves [7], which support evaluation of model performance as a function of misclassification costs.”
A systematic study of the class imbalance problem in convolutional neural networks (2017, methodology): “For example, cost curves [100, 101], regret graphs introduced by Hilden and Glasziou [102], performance graphs used by Turney [103], Loss Difference Plots, and LC index [104].”
Learning from Imbalanced Data (2009, background): “Another shortcoming of ROC curves is that they lack the ability to provide confidence intervals on a classifier’s performance and are unable to infer the statistical significance of different classifiers’ performance [128], [129].”
Information, Divergence and Risk for Binary Experiments (2009, methodology): “In this paper we draw connections between Receiver Operating Characteristic (ROC) curves, (Fawcett, 2006, 2004; Flach, 2003; Flach and Wu, 2005; Maxion and Roberts, 2004) the Area Under ROC Curve (AUC), (Cortes and Mohri, 2004; Hand, 2008; Hand and Till, 2001; Hanley and McNeil, 1982) and Cost Curve”
EnsembleMatrix: interactive visualization to support machine learning with multiple classifiers (2009, neutral): “More general techniques that apply across algorithm types include ROC and Cost curves [8], which support evaluation of model performance as a function of misclassification CHI 2009 ~ Visualization 2 April 8th, 2009 ~ Boston, MA, USA”