**4. Experimental Evaluation**

The proposed solution identifies context-aware sociability patterns using incremental unsupervised learning. Consequently, applying metrics commonly used to evaluate learning models is challenging because there is no ground truth available to compare results. By considering this, we evaluate the proposed solution's feature to consistently recognize sociability patterns for modeling social routine and behavior changes.

In [20], we performed the following: (i) a comparison of the similarity between the sociability patterns identified with the intervals recognized by Gaussian Mixture Models (GMMs); (ii) an analysis of the contribution of context-aware sociability patterns to understand the monitored individual's social routine. Here, we performed a more in-depth evaluation to recognize the proposed solution ability to identify context-sensitive sociability patterns that model social behavior and detect behavioral changes.
