Fig. 4
Results of regression analysis. (a) Performance of each regression model is presented as a function of feature combination cases sorted on R-squared value. (b–f) The best case of each model is shown. Dots represent the pain scores of the subjects predicted by each model, and the black line represents the actual pain scores of the subjects. The correlation coefficients of each model are presented in parentheses, and ch denotes the statistical chance level of the best features. The chance levels were computed by averaging the \({R}^{2}\) values over 1000 random shuffling of labels. The asterisk (*) indicates statistical significance meaning that the performance is different from the statistical chance level, as confirmed using a one-sample t-test (p < 0.01).
