Human body carbohydrate and fat modeling by using artificial neural networks

In this study, cardio pulmonary exercise tests (CPET) data, 6-minute walk test data and anthropometric data were modeled using artificial neural networks. 44 sedentary male individuals ages ranging from 20 to 30 were participated in the study. Following anthropometric measurements all Participants were applied to 4 different exercise tests which including maximal cardiopulmonary exercise test, highest fat oxidation rate detection test (Fatmax), 40 minutes walking and 6 minutes walking test. In artificial neural network models, it is aimed to estimate the data of CPET, by using more simpler and easily applied 6 minutes walk test and anthropometric measurement data which required serious physical infrastructure, advanced laboratory equipment and trained personnel in the past. In the ANN models, the correlation between the predicted values of the training, verification and test sets and the actual values was found to be 0.98 and above. 40 minutes walking test data parameters including CHO and Fat (kkal / day) were successfully estimated in the results as well as by using parameters from CPET Fatmaks test and CPET 40 minutes walking test data. In cases where CPET are not available, 6-minute walk test and anthropometric measurements, which are one of the simpler and more economical alternatives that can be used at the primary health care level, are modeled.