Epidemiological Feature Analysis And Prediction System Of ARIMA-PGXGBoost Model Based On Wavelet Transform

In the field of epidemiology,the incidence of infectious diseases often has periodic and seasonal peak characteristics in annual units.In modern infectious disease control strategies,the most fundamental approach is to increase the detection rate of patients and effectively curb the infection.Therefore,it is of great significance to accurately predict the number of infectious diseases.Hand-foot-mouth disease is a common infectious disease in infants and young children,it has the typical characteristics of infectious diseases and can be used as an example to construct an epidemiological feature analysis and prediction system.Existing hand-foot-mouth disease analysis methods include traditional statistical analysis models and new machine learning models,as well as combined models based on these single models.The combined model has better prediction effects than the single model but has limited improvement.This research first considered the optimization of Integrated learning algorithm based on decision tree XGBoost.Aiming at the difficulty of tuning parameters of the ordinary XGBoost prediction system,particle swarm optimization algorithm PSO and genetic algorithm GA are introduced to construct an infectious disease prediction system based on PSO-GA-XGBoost.The genetic algorithm optimizes the parameters of the XGBoost model based on the idea of survival of the fittest,but it has the disadvantage of easily falling into local minima,so particle swarm optimization algorithm is introduced to overcome local minima and perform better search ability to optimize the combination of parameters and achieve better model prediction accuracy.Experiments verify the superiority of PSO-GA-XGBoost compared to the ordinary XGBoost model.The sequence of hand-foot-mouth disease is often mixed with the trend part and the nonlinear part,which may cause the incomplete feature extraction of PGXGBoost in the modeling process.The wavelet transform can separate the trend part and the nonlinear part to a certain extent,so we consider introducing wavelet transform to decompose the sequence of hand-foot-mouth disease.At the same time,the statistical prediction method ARIMA is used to process the trend sequence decomposed by wavelet transform,PGXGBoost is used to process the nonlinear sequence,and ARIMA and PGXGBoost are good at processing the trend and the nonlinear part respectively.Based on these theories,this study constructed an ARIMA-PGXGBoost infectious disease prediction system based on wavelet transform.Experiments have verified that it has very good predictive performance.It provides good theoretical support for formulating prevention measures and strategies,and it has very good application prospects for the analysis and prediction of the characteristics of infectious diseases.