Statistical Analysis Of Longitudinal Count Data Based On The Random Effects Model

This paper mainly discusses the statistical analysis of the longitudinal count data arising from recurrent event.Longitudinal count data,also known as recurrence event data,refers to the data obtained by continuous observation of multiple study individuals and recording the time when a certain event repeatedly occurred.Compared with normal continuous Longitudinal data,the main difference of Longitudinal count data is that the observed variable is no longer a continuous variable,but corresponds to a counting process.Longitudinal count data widely exists in the field of social sciences,biology and economics,then the statistical analysis of longitudinal count data has important theory value and practical significance.In longitudinal count data,recurrence event there is dependence between time to relapse for many times,so this paper will revolve around this topic to discuss.establish the random effects model for simple count data,and the fixed effects model and random effects model for longitudinal count data.we establish the maximum likelihood estimation procedure to estimate the unknown parameters in the model and prove large sample properties of the estimates.Numerical simulations are also conducted combined with real data analysis for a set of data on the times to development of mammary tumors.Finally,we summarize the established models and statistical inferences,and proposes some ongoing work for further research..