A First-order Integer Autoregressive Model Considering the Influence of Covariates on Mixed Probabilities
Selecting appropriate counting data statistical methods can help solve data analysis and prediction problems in the era of big data.The study utilizes a logistic structure to integrate a First-Order integer-valued autoregressive model that considers the influence of covariates on mixing probability,and combines the unknown parameters in the conditional maximum likelihood estimation model.The research results show that as the sample size increases,the bias,standard deviation,and mean square error of the model parameter estimation decrease,indicating that the model estimators are consistent.Under the condition of parameter combination of(0.6,0.5,1)and sample size of 800,Model A exhibits low bias,low standard deviation,and low mean square error,and the conditional maximum likelihood estimator follows a normal distribution trend.When comparing First-Order integer autoregressive models,the results based on the Chichi information criterion and Bayesian information criterion indicate that the research model has better fitting performance and is more helpful for for the prediction and analysis of computational data.
first-order integer autoregressionlogistic structurecovariancemixed probabilityconditional maximum likelihood estimationparameter estimation
NETL
NSTL
万方数据
