Research on human body composition prediction model based on Akaike Information Criterion and improved entropy method

Abstract

The prediction model of existing human body composition based on measured bioelectricity has problems that include redundant influence factors and low prediction accuracy. To address these problems, this paper put forward a human body composition prediction model based on Akaike Information Criterion (AIC) and improved entropy method. First, combining with the AIC information principle, we selected a set of characteristic parameters from human physiological arguments, and constructed the human body composition prediction model; Second, improved entropy method was used to solve the unknown coefficients in predictive model, then worked out prediction model of human body composition; Finally, a comparative analysis experiment of the prediction model and the actual measurement data was designed, and the data were sampled by InBody770 body composition instrument. Experimental results showed that a good correlation existed between the model predictions data and the actual measurements, this study provided a theoretical basis for the model and analysis of human body composition.