Artificial Intelligence Based Estimation of Individuals’ Daily Energy Requirements with Anthropometric Measurements and Demographic Information

Energy must be taken regularly for the body to continue its life healthily. Age, gender, physical activity, height, and weight determine the energy required to sustain life. There are various methods used to measure energy expenditure. These methods; methods such as direct calorimetry, gold standard indirect calorimetry, respiratory coefficient, and double effect, can be counted. These methods can be counted as; direct calorimetry, gold standard indirect calorimetry, respiratory coefficient, and double effect. Therefore, newmethods are needed. This study proposes anthropometric and demographic information on the individual and artificial intelligence-based energy needs estimation. Artificial intelligence models were created by selecting 14 characteristics of individuals with the help of selection. Linear, Gaussian Process Regression (GPR), Neural Network (NN), Support Vector Machine (SVM), and ensemble methods were used as artificial intelligence models and developed gender-based models for increased performance. The best model performance was determined in the models developed for all individuals as RMSE = 1.6, R2 = 1 MSE = 2.58, and MAE = 0.74 with a GPR-based 50% feature. 50% means that 50% of the features are used. In the models developed for men, the best model performance was determined as RMSE = 0, R2 = 1 MSE = 0, and MAE = 0 with a linear-based 50% feature. In the models developed for women, the best model performance was determined as RMSE = 0, R2 = 1 MSE = 0, and MAE = 0 with a linear-based 75% feature. According to the results, anthropometric and demographic information and artificial intelligence-based energy requirement can be calculated accurately for all individuals or based on gender.