Adaptive Cost-Sensitive Learning in Neural Networks for Misclassification Cost Problems
This addresses cost-sensitive classification problems where different misclassification errors have varying consequences, though it appears incremental as an adaptive extension to existing methods.
The paper tackles misclassification cost problems by developing an adaptive learning algorithm (AdaCSL) that adjusts the loss function to address local training-test class distribution mismatches, showing it yields better cost results on both class-imbalanced and balanced binary classification datasets compared to alternatives.
We design a new adaptive learning algorithm for misclassification cost problems that attempt to reduce the cost of misclassified instances derived from the consequences of various errors. Our algorithm (adaptive cost sensitive learning - AdaCSL) adaptively adjusts the loss function such that the classifier bridges the difference between the class distributions between subgroups of samples in the training and test data sets with similar predicted probabilities (i.e., local training-test class distribution mismatch). We provide some theoretical performance guarantees on the proposed algorithm and present empirical evidence that a deep neural network used with the proposed AdaCSL algorithm yields better cost results on several binary classification data sets that have class-imbalanced and class-balanced distributions compared to other alternative approaches.