The time-series data analysis described here is based on the simple idea that the accumulation of the effects of lifestyle events, such as ingestion and exercise, could affect personal health with some delay. The delay may reflect complex bio-reactions such as those of metabolism in a human body. In the analysis, the accumulation of the effects of lifestyle events is represented by a summation of daily lifestyle data whose time-series correlation to variations of health data is examined (healthcare-data-mining). The concept of weighting is introduced for the summation of daily lifestyle data. As a result, it is suggested that the nature of personal health could be represented by a weighting pattern characterized by a small number of parameters.

Correlations between energy expenditure/supply and body-fat percentage were studied using personally stored daily time-series data. The weighting patterns for the summation of daily time-series energy expenditure and supply data giving the maximum correlations with the variation of daily body-fat percentage data were obtained. The weighting patterns can be expressed by two parameters whose combination is considered to characterize the nature of personal health. The combination of the parameters for a subject was found to show a significant bias in the frequency distribution, independent of season and aging, for the term of seven years, and the combination of the parameters of 20 other subjects showed a tendency to divide into two types.