论文部分内容阅读
选用符合林火发生数据结构的Poisson和零膨胀Poisson(ZIP)模型对大兴安岭林区1980—2005年间林火发生与气象因素关系进行建模分析,并与普通最小二乘回归(ordinary least squares,OLS)方法的结果进行了对比分析.结果表明:OLS模型对研究区域林火发生与气象因素关系的拟合结果较差(R2=0.215);Poisson和ZIP模型的拟合效果较好,具有较好的火灾次数预测能力,且ZIP模型的预测能力高于Poisson模型.运用AIC和Vuong检验方法对Poisson和ZIP模型的拟合水平进行进一步检验,表明ZIP模型的数据拟合度优于Poisson模型.
The Poisson and zero-expansion Poisson (ZIP) models were used to analyze the relationship between forest fire occurrence and meteorological factors in Daxing’an Mountains from 1980 to 2005. The results were compared with ordinary least squares regression (OLS) The results show that the OLS model has a poor fitting result (R2 = 0.215) for the relationship between forest fire occurrence and meteorological factors, and the Poisson and ZIP models have good fitting results and good results , And the predictive ability of ZIP model is higher than that of Poisson model.Using AIC and Vuong test to further verify the fit of Poisson and ZIP models, it shows that the fitting degree of ZIP model is better than that of Poisson model.