Figure 2 | Scientific Reports

Figure 2

From: Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience

Figure 2

[ANS Data] Results from supervised and unsupervised analyses of ANS data. (a) Mean ± SEM classification accuracy for the top three emotion words used per participant. The red line represents the mean of the null distribution for chance performance of the group. (b) Distribution of participants by number of discovered physiology clusters resulting from DP-GMM. (c) Correspondence between the discovered clusters and the top three participant-specific emotion categories for three representative example participants with four, five, and seven clusters respectively (corresponding to Participant IDs 40, 32, and 26 in part a, with respective classification accuracy levels of 82.26%, 74.27%, and 73.80%). Bars represent the proportion of participant-generated emotion category per cluster. The total proportion of categories in each cluster sums to 1. The total number of emotion category labels used by example participants 40, 32, and 26 were 26, 17, and 9, respectively. The mixing proportion \(\pi _k\) reported below each cluster is the probability that an observation comes from that cluster, which is representative of the size of the cluster.

Back to article page