Causality acquisition from a large sample of vital data for a weight change prediction system

In today's society, the awareness of healthcare and public health has been increasing. Reflecting this trend, there have recently been many reports aimed at the development of an intelligible framework that does not focus on merely medical issues but rather on how to efficiently utilize various human health-related data with IT systems. This study proposes an approach to modeling causality-based dynamics in human health-related parameters. A prediction system for weight change due to, apparently, physical exercises and subjective amounts of meals per day is developed. The approach builds on the Principal Component Analysis (PCA) technique to detect dependencies among a number of health variables collected through the study. Results obtained through the analysis are used to derive a simple model that allows one to predict human weight changes with a high accuracy. The model can be used in various health-monitoring applications.