收稿日期: 2009-11-02
录用日期: 2010-01-13
网络出版日期: 2010-05-01
Relationship between forest lighting fire occurrence and weather factors in Daxing’an Mountains based on negative binomial model and zero-inflated negative binomial models
Received date: 2009-11-02
Accepted date: 2010-01-13
Online published: 2010-05-01
雷击火的发生与气象因子之间存在着密切的关系。该文选用符合大兴安岭地区林火发生数据结构的负二项(negative binomial, NB)和零膨胀负二项(zero-inflated negative binomial, ZINB)两种模型对大兴安岭林区1980-2005年间雷击火的发生与气象因素间的关系进行建模分析, 并与以往研究中所使用的最小二乘(OLS)回归方法相对比。使用SAS和R-Project统计软件进行模型拟合运算, 计算得出模型各参数。结果表明, NB和ZINB模型对数据拟合较好, 模型内各气象因子显著性水平较高, 对雷击火发生次数均具有较好的预测能力。运用AIC和Vuong等检验方法, 进一步比较了NB和ZINB模型对数据的拟合水平以及模型预测水平, 结果表明ZINB模型无论在数据拟合还是模型预测上都要优于NB模型。提出了大兴安岭地区林火发生与气象因子关系的最优模型。
郭福涛, 胡海清, 金森, 马志海, 张扬 . 基于负二项和零膨胀负二项回归模型的大兴安岭地区雷击火与气象因素的关系[J]. 植物生态学报, 2010 , 34(5) : 571 -577 . DOI: 10.3773/j.issn.1005-264x.2010.05.011
Aims Much research has been carried out on the relationship between forest fire occurrence and weather factors by use of modeling in recent years. However, the data organization used in past research can not satisfy the requirements of models well. Our aims are to determine the regression model that best fits the forest fire data set and provides a new model theory for research on forest fire and its influencing factors in order to forecast lighting fire occurrence.
Methods We used negative binomial (NB) and zero-inflated negative binomial (ZINB) models to describe the relationship between lighting fire occurrence and weather factors in the Daxing’an Mountains for 1980-2005 using SAS 9.1 version and R-Project statistic software and comparing results from these models by use of AIC and Vuong methods.
Important findings Both NB and ZINB models produced results with high significance for each weather factor. Comparison of the two models according to AIC, Vuong and other methods showed that the fitting ability and predictive power of ZINB model are better than those of the NB model. The advantage was also found when we compared the modeling results with Ordinary Least Squares. Then we obtained the best model for the relationship between lighting fire and weather factors.
| [1] | Akaike H (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19, 716-723. |
| [2] | Cameron AC, Trivedi PK (1998). Regression Analysis of Count Data. Cambridge University Press, Cambridge, UK. |
| [3] | Dean CB, Lawless JF (1989). Tests for detecting over-dispersion in Poisson regression models. Journal of the American Statistical Association, 84, 467-472. |
| [4] | Hall DB (2000). Zero-inflated Poisson and binomial regression with random effects: a case study. Biometrics, 56, 1030-1039. |
| [5] | Jesper R, Igor R (2006). A note on estimation of intensities of fire ignitions with incomplete data. Fire Safety Journal, 41, 399-405. |
| [6] | Kelvin KW, Wang YK, Lee AH (2003). Zero-inflated negative binomial mixed regression modeling of over-dispersed count data with extra zeros. Biometrical Journal, 45, 437-452. |
| [7] | Krawchuk MA, Cumming SG, Flannigan MD (2009). Predicted changes in fire weather suggest increases in lightning fire initiation and future area burned in the mixedwood boreal forest. Climatic Change, 92, 83-97. |
| [8] | Lambert D (1992). Zero-inflated Poisson regression with an application to defects in manufacturing. Technometrics, 34, 1-14. |
| [9] | Martin R, John H, Clarice GB (2001). A score test for testing a zero-inflated poisson regression model against zero- inflated negative binomial alternatives, Biometrics, 57, 219-223. |
| [10] | Miao BQ (缪柏其), Wei J (韦剑), Song WG (宋卫国), Tan CC (谭常春) (2008). Logistic and ZIP regression model for forest fire data. Fire Safety Science (火灾科学), 17, 143-149. (in Chinese with English abstract) |
| [11] | Miller AJ (1992). Algorithm AS 274: Least squares routines to supplement those of gentleman. Journal of the Royal Statistical Society, Series C (Applied Statistics), 41, 458-478. |
| [12] | Mullahy J (1986). Specification and testing of some modified count data models. Journal of Econometrics, 33, 341-365. |
| [13] | Nocedal J, Wright S (1999). Numerical Optimization. Springer-Verlag. USA. |
| [14] | Quintanilha JA, Ho LL (2006). Analyzing wildfire threat counts using a negative binomial regression model. Environmetrics, 17, 529-538. |
| [15] | van den Broek J (1995). A score test for zero zero-inflation in a Poisson distribution. Biometrics, 51, 738-743. |
| [16] | Vuong QH (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57, 307-333. |
/
| 〈 |
|
〉 |
