Probabilistic Latent Semantic Sturucture Modeling Based on The Large Scale Elderly Living Activities Survey

Abstract

In Japan, there are many social problems caused by super-aged society. So we need to resolve these problems by using ICT. In addition, it is also necessary to actively use of big data in healthcare. In this study, we analyzed the elderly questionnaire to investigate relations between activity and subjective happiness level in this analysis. We used probabilistic Latent Semantic Analysis and Bayesian network. We found variables related to elderly heterogeneity and behavioral character. These findings make it possible to simulate elderly's activity and provide new insight into elderly's subjective level of happiness.