基于复合泊松过程与时间序列模型的新疆森林火灾特征预测分析研究  

Study on Prediction and Analysis of Forest Fire Characteristics in Xinjiang Based on Compound Poisson Process and Time Series Model

作　　者：王瑶 卢芸潇 刘淼 Wang Yao;Lu Yunxiao;Liu Miao(College of Mathematics and Statistics,Yili Normal University,Yining,Xinjiang 835000,China;Institute of Applied Mathematics,Yili Normal University,Yining,Xinjiang 835000,China)

机构地区：[1]伊犁师范大学数学与统计学院,新疆伊宁835000 [2]伊犁师范大学应用数学研究所,新疆伊宁835000

出　　处：《伊犁师范大学学报（自然科学版）》2023年第2期16-24,共9页Journal of Yili Normal University：Natural Science Edition

基　　金：新疆维吾尔自治区自然科学基金项目(2021D01C462);伊犁师范大学“学实高层次人才岗位”项目(YSXSJS22006)。

摘　　要：基于时间序列ARIMA模型对新疆2003~2019年期间森林火灾燃烧面积的相关数据进行实证预测分析,进一步提出运用复合泊松过程可以对新疆每年森林火灾燃烧次数发生概率模拟.预测结果显示:2020~2021年平均燃烧面积约为40 hm~2,以及新疆年均森林火灾预期次数达到37次.可见,复合泊松模型可以作为新疆森林火灾燃烧总面积概率拟合的一个很好的工具.Based on the ARIMA model to predict the relevant data of forest fires in Xinjiang from 2003 to 2019,this paper further proposes that the compound Poisson process can simulate the annual combustion area and combustion frequency probability of forest fires in Xinjiang.The results predict that the average burning area from 2020 to 2021 is 40 hm~2;and that the average annual expected number of forest fires in Xinjiang is 36.88.

关 键 词：复合泊松过程模型 差分自回归移动平均模型 森林火灾 

分 类 号：O211.6[理学—概率论与数理统计]