k阶最近邻距离混合点过程分解模型的适用性判断条件的修正  

Rectification on Applicability of Mixed Point Process Decomposition

作　　者：杨萍[1] 侯威[2] 封国林[2] 

机构地区：[1]中国气象局北京城市气象研究所,北京100089 [2]国家气候中心,北京100081

出　　处：《计算物理》2011年第6期875-882,共8页Chinese Journal of Computational Physics

基　　金：国家自然科学基金(41005043;41175067);全球变化研究国家重大科学研究计划(2012CB955901);科技部支撑项目(2007BAC03A01和2007BAC29B01);公益性行业(气象)科研专项(GYHY201106016)资助项目

摘　　要：对k阶最近邻距离混合点过程分解模型的判断条件进行修正,使该模型的适用范围更接近于真实.数值试验发现,参数R的理论值较大或者较小均不影响模型适用范围的判断条件.对适用范围的理论临界值可能产生影响的主要是该临界值附近的R.分析临界值及其附近R与相应的理论值之间的差异,结合计算值,对所确定的临界值进行修正.结果发现:修正后的临界值一般都小于原临界值;无论临界值是否修正,点的个数对临界值的影响相似,即数据点数目较少时,临界值较大,数据点数目较大时,临界值较小;临界状态下,k的初始有效值基本随着数据点数目的增加而有所增加.A mixed point process decomposition based on k-th nearest distance is introduced.Disadvantage in former researches is revealed and resovled.It is shown that the threshold of applicability judgment is not influenced as R is larger or smaller than critical value.Primary influences show as R is close to the theoretical calculation.Difference between theoretical and calculated R is analyzed.And numerical experiment is done.It is noticed that whether the critical value is revised or not,it depends on the number of point with almost a negative correlativity.In other words,the critical value is higher as sample size is small but is lower as sample size is large.As R lies in critical conditions,the initialized valid value of k increases with increasing of sample size.

关 键 词：丛集点 k阶最近邻距离 分布参数比值 修正 

分 类 号：P468[天文地球—大气科学及气象学]