EM Algorithm Parameter Estimation Problem Based On Group Testing

In the process of biomedical detection,how to detect a disease efficiently,accurately and economically and estimate the overall prevalence rate is a topic worth studying.In the detection of a certain disease,it is common to use their blood or samples containing individual information for detection.However,in the case of large sample size,grouping detection is a very efficient and convenient method for the detection of individuals.Grouping detection is very practical in large-scale detection because its principle is simple,easy to understand,easy to operate and can effectively reduce the number of tests,reduce the test cost and improve the detection efficiency.This paper mainly studies the problem of parameter estimation in the case of two group sizes.In this paper,we mainly study the mathematical model in which the detection samples are interfered by random error terms.Considering the size of two groups,we establish a mathematical model based on the obtained data and study the estimation of parameters in the model.In group detection,the detected individuals are interfered by random errors and the size of multiple groups is considered.According to the incomplete data obtained from the observation,an implicit variable is introduced,namely the number of positive individuals in each group in the grouping detection.Then,the joint density function is obtained from the complete data,the maximum likelihood estimation function is calculated,and the estimated value of parameters in the model is calculated by the EM algorithm.Then better parameter estimation is obtained by simulation,and the accuracy of parameter estimation is improved.