Clustering using principal component analysis applied to autonomy–disability of elderly people

The aim of this paper is to find feature-patterns related to the autonomy–disability level of elderly people living in nursing homes. These levels correspond to profiles based on the people's ability to perform activities of daily living like being able to wash, dress and move. To achieve this aim, an unsupervised approach is used. In this article, we propose a new clustering approach based on principal component analysis (PCA) to better approximate clusters. We want to automatically find categories or groups of residents based on their degree of autonomy–disability. All residents in a group have similar patterns. The main function of PCA is to explore the links between variables and the similarities between examples (individuals). The proposed algorithm uses the PCA technique to direct the determination of the clusters with self-organizing partitions by using the Euclidian distance. The study was carried out in close collaboration with the French cooperative health organization called the “Mutualité Française de la Loire”. The quantitative data arises from the databases of four different nursing homes located in the city of Saint-Etienne in France. The study concerns 2271 observations of dependence evaluations corresponding to 628 residents.