Evaluation of error- and correlation-based loss functions for multitask learning dimensional speech emotion recognition

Abstract

Abstract The choice of a loss function is a critical part in machine learning. This paper evaluates two different loss functions commonly used in regression-task dimensional speech emotion recognition — error-based and correlation-based loss functions. We found that using correlation-based loss function with concordance correlation coefficient (CCC) loss resulted in better performance than error-based loss functions with mean squared error (MSE) and mean absolute error (MAE). The evaluations were measured in averaged CCC among three emotional attributes. The results are consistent with two input feature sets and two datasets. The scatter plots of test prediction by those two loss functions also confirmed the results measured by CCC scores.