Data-driven Spatiotemporal Measurment Model And Infereence Methods For Infectious Diseases

Although with the improvement of the social level and the development of science and technology,the medical level has been significantly improved,but human beings are still very vulnerable in the face of infectious diseases.How to scientifically monitor and control infectious diseases is a big problem facing mankind.As early as the 18 th century,scholars established dynamic models to characterize and study the spread of infectious diseases.However,infectious disease models based on differential equations are difficult to quantify the impact of various dynamic factors that change over time(such as climate change)on the epidemic of infectious diseases.In recent years,with the continuous development of information technology,data-driven infectious disease models have achieved unprecedented development.However,the data-driven infectious disease model has the characteristics of non-linearity,high dimensionality,and multi-parameters,which puts forward higher requirements on how to dynamically reason about model parameters.This article will take malaria and new coronary pneumonia as examples to integrate multi-source data such as environment,weather,and population movement to build a data-driven infectious disease model.Furthermore,based on the time series data of the epidemic,the unknown parameters in the model are estimated and inferred by an improved Markov chain-Monte Carlo algorithm.First,we proposed a state-space model based on time series data to realize the trend prediction of malaria epidemics.By comprehensively considering factors such as environment,weather,media energy,and population,a data-driven state space model is constructed.Furthermore,by linking the malaria transmission model with the epidemic observation data,a particle Markov chain-Monte Carlo algorithm was proposed to scientifically evaluate the model parameters(such as transmission efficiency,recovery rate,etc.)that affect malaria transmission.Based on the model parameters and the hidden state process,we propose a step forward prediction method to predict the trend of malaria transmission.The results of malaria epidemic analysis in Tengchong,Yunnan,my country show that the data-driven state space model proposed in this paper achieves higher prediction accuracy than traditional statistical prediction methods.Secondly,we proposed a propagation model based on population movement and conducted a retrospective analysis of the new crown pneumonia epidemic in my country.Specifically,based on the SEIR model and the particle Markov chain-Monte Carlo algorithm,we estimated the illness time of Wuhan case zero.On this basis,considering the impact of population movement on the spread of new coronary pneumonia in other cities in Hubei Province,we have established a new coronary pneumonia transmission model based on population movement.By studying the impact of different lockdown dates on the number of confirmed patients with new coronary pneumonia,the key role of the lockdown in controlling the spread of new coronary pneumonia was quantitatively analyzed.In summary,this paper constructs a data-driven propagation model by fusing various influencing factors that affect the epidemic situation of infectious diseases,and proposes a particle Markov chain-Monte Carlo algorithm to estimate the model parameters.The model and calculation method proposed in this paper can be used to predict the spreading trend of infectious disease epidemics,assess the time of illness of the zero case of infectious diseases,and analyze the impact of population movement on the spread of infectious diseases in time and space,in order to achieve scientific prevention of infectious disease epidemics.Provide scientific basis for control.