利用吉林省汪清林业局金沟岭林场落叶松林分连续观测数据,分别利用Poisson回归模型、负二项模型、零膨胀模型和Hurdle模型拟合林木枯损株数,并通过AIC值以及Vuong检验对这些模型进行详细分析比较.结果表明:Poisson回归模型不适用于模拟林木枯损株数,负二项回归模型相对于Poisson回归模型比较适用；但是对于零枯损过多的数据,这2类模型拟合效果较差.零膨胀模型和Hurdle模型对这类数据有很好的解决办法,其中,零膨胀负二项模型和Hurdle-NB模型拟合效果优于其他几种模型,且Hurdle-NB模型略好于零膨胀负二项模型.%Stand mortality is a very important variable for describing the stand characters. Based on the stand mortality data from permanent plots of Larix spp. in Wangqing Forest Farm, Poisson model, negative binomial model, zero-inflated model and Hurdle model were introduced to model the stand mortality stems. And the best model was selected through AIC and Vuong test. Results showed that; Poisson model was not suitable for stand mortality, and negative was superior to the Poisson model. But both of them were not competent for the over-dispersion data of stand mortality. Zero-inflated model and Hurdle model were fitted into the data. Additionally, zero-inflated negative binomial model(ZINB) and Hurdle-NB model outperformed than other models. Furthermore, The Hurdle-NB model was a little better than ZINB model.
展开▼
