A Bayesian Approach to Cluster Analysis of Ekman 60-Faces Test in Degenerative Diseases Patients’ Cohort
摘要
This work presents an application of the Bayesian Bagged Clustering (BBC) algorithm for the analysis of Ekman 60-Faces Test results in a population of patients affected by neurodegenerative diseases. The hypothesis that the type of emotions recognized or mistaken can be informative on the impairments of the patient and potentially helpful for the differential diagnosis is tested. The advantages of the application of advanced clustering techniques are shown in terms of choice of the number of clusters, cluster assignments and information retrieval on the data. The results obtained suggest that there is a weak association between the type of emotion mistaken and the diagnosis of the subjects, and that other clinical characteristics of the subjects should be taken into account to understand the results.