Simulation Analysis of L-M Optimization Algorithm for Weight Loss under BP Network Model

Abstract:
In recent years, the prevalence of obese patients in China and the world has been increasing. Obese patients lack selfesteem and self-confidence in everyday activities, such as work and study. Apart from its adverse socio-psychological impact, obesity is well known to be also harmful to the human body. High levels of obesity lead to the accumulation of excessive quantities of subcutaneous fat, which can easily induce hyperlipidemia. Furthermore, obesity causes hyperglycemia and hyperuricemia. Certain negative effects are also exerted on male and female reproductive function, eventually leading to infertility. In this paper, a mathematical model was built based on relevant knowledge to effectively explore approaches to achieving healthy weight loss. Based on BP neural network, a practical weight loss model was constructed. In the processes of simulation and prediction, this study applied Bayesian regularization, L-M optimization, and momentum gradient descent algorithms. Then, in accordance with the algorithm, a formula was utilized to predict the whole weight loss process, and the actual result and error values of each algorithm were emphatically explored. Our present findings indicate that the L-M optimization algorithm can effectively and precisely simulate and predict the weight loss process, achieving the best simulation and prediction effects.