Extended Data Fig. 6: Comparison between AIC-based clustering and unsupervised clustering approaches. | Nature Mental Health

Extended Data Fig. 6: Comparison between AIC-based clustering and unsupervised clustering approaches.

From: Individual differences in autism-like traits are associated with reduced goal emulation in a computational model of observational learning

Extended Data Fig. 6

The ‘Lowest AIC Approach’ classifies individuals using a fixed boundary. The ‘Unsupervised Clustering’ represents our current data-driven subject clustering approach, which does not have a hard boundary between groups, but clusters individuals based on the cosine distance of dimension values in a high-dimensional space. The top right (green) cluster is a separate subject group (Group 3) in the latter approach but is split into two other groups (Group 1, Group 2) in the former approach.

Back to article page