Parameter estimation and statistical testing of 0-k inflation count regression models have always been a hot issue in data analysis.The 0-k inflation regression model is reset with the help of latent variables,and the EM algorithm and Fisher information matrix are used to give the point and interval estimates of the parameters.Based on the Wald statistic and the LR statistic,a statistical test is carried out to test whether the 0-k inflation distribution will degenerate.Take the 0-k inflated Poisson regression model as an example for practical application.Research shows:When the EM algorithm is used for parameter estimation,the convergence speed is fast and the estimation results are robust.Different 0-k inflated Poisson distributions fit the original data differently.Setting k to a non-dilated value may lead to biased fitting results.
关键词
0-k膨胀/回归模型/参数估计/假设检验/泊松分布
Key words
0-k inflation/regression models/parameter estimation/hypothesis testing/Poisson distribution
维普
万方数据