The identification for time-varying parameter and noise of tuberculosis with seasonal changes based on data-behavior-driven system

Time-varying and seasonal parameter inversion in the mathematical model of infectious diseases and uncertainty quantization based on actual data have great significance for real quantitative transmission process. In this study, the behavior-driven mathematical model of infectious diseases and the data-driven parameter identification method are combined to quantify the transmission law of tuberculosis (TB). To begin with, according to the characteristics of TB transmission, the TS-SID model with time-varying is established. Then, the improved identification algorithm is proposed to track the fluctuation of disease infection rate and mortality rate considering the seasonal influence. Meanwhile, focusing on the influence of noise on the spread of diseases, noise reduction and uncertain quantization are carried out on the data to identify the noise distribution. In addition, predict the denoised sequence and superimpose the noise distribution, which can improve the rationality of prediction. Finally, the numerical comparison shows that seasonal time-varying tracking is good for grasping and predicting the disease evolution.