Abstract:Intelligent computer systems aim to help humans in making decisions. Many practical decision-making problems are classification problems in their nature, but standard classification algorithms often not applicable since they assume balanced distribution of classes and constant misclassification costs. From this point of view, algorithms that consider the cost of decisions are essential since they are more consistent with the requirements of real life. These algorithms generate decisions that directly optimize parameters valuable for business, for example, the costs savings. But despite on practical value of cost-sensitive algorithms, the little number of works study this problem concentrating mainly on the case when the cost of a classifier error is constant and does not depend on a specific example. However, many real-world classification tasks are example-dependent cost-sensitive (ECS), where the costs of misclassification vary between examples and not only within classes. Existing methods of ECS learning include just modifications of the simplest models of machine learning (naive Bayes, logistic regression, decision tree). These models produce promising results, but there is a need for further improvement in performance that can be achieved by using gradient-based ensemble methods. To break this gap, we present the ECS generalization of AdaBoost. We study three models which differ by the ways to introduce cost into the loss function: inside the exponent, outside the exponent, and both inside and outside the exponent. The results of the experiments on three synthetic and two real datasets (bank marketing and insurance fraud) show that example-dependent cost-sensitive modifications of AdaBoost outperform other known models. Empirical results also show that critical factors influencing the choice of the model are not only the distribution of features, which is typical for cost-insensitive and class-dependent cost-sensitive problems but also the distribution of costs. Next, since the outputs of AdaBoost are not well calibrated posterior probabilities, we check three approaches to calibration of classifier scores: Platt scaling, isotonic regression, and ROC modification. The results show that calibration not only significantly improves the performance of specific ECS models but allows making better capabilities of original AdaBoost. Obtained results provide new insight regarding the behavior of the cost-sensitive model from a theoretical point of view and prove that the presented approach can significantly improve the practical design of intelligent systems.

Optimization of Selective Ensemble for Cost-Sensitive Classification: an Empirical Study

Support Vector Machines Ensemble With Optimizing Weights By Genetic Algorithm

Cost Sensitive Support Vector Machines

SVM-Based Cost-sensitive Classification Algorithm with Error Cost and Class-dependent Reject Cost

An adaptive Cost-sensitive Classifier

Cost-sensitive probability for weighted voting in an ensemble model for multi-class classification problems

Decision-Theoretic Rough Set and Cost-Sensitive Classification

A Decision-Theoretic Rough Set Approach To Multi-Class Cost-Sensitive Classification

Cost-sensitive selection of variables by ensemble of model sequences

Adaptive Cost-Sensitive Online Classification

OptimizingEnsemble Learning to Reduce Misclassification Costs in Credit Risk Scorecards

Self-adaptive cost weights-based support vector machine cost-sensitive ensemble for imbalanced data classification

An adaptive cost-sensitive learning approach in neural networks to minimize local training–test class distributions mismatch

Towards cost-sensitive learning for real-world applications

Cost-Sensitive Classification Using Decision Trees, Boosting and MetaCost

An improved selective ensemble learning approach in enabling load classification considering base classifier redundancy and class imbalance

Financial Predictions Using Cost Sensitive Neural Networks for Multi-Class Learning

The Influence of Class Imbalance on Cost-Sensitive Learning: an Empirical Study.

Example-dependent cost-sensitive adaptive boosting