Comparative Study on the Powers Based on F*, G* and J*

基于F*，G*和J*的功效比较研究

10.6338/JDA.200710_2(5).0009

吴晓坤(Xiao-Kun Wu)；吴刘仓(Liu-Cang Wu)；李会琼(Hui-Qiong Li)；唐年胜(Nian-Sheng Tang)

功效 ； 蒙特卡洛检验 ； 点过程 ； power ； Monte Carlo test ； point processes

Journal of Data Analysis

2卷5期（2007 / 10 / 01）

135 - 146

英文

检验统计量在点过程中有着非常重要的地位，它可用于判断点模型的类别。陈（2003）提出了三个新的统计量F*，G*和J*，而且证明了G*和J*在大样本的情形有非常好的性质。但陈没有比较研究它们的功效。本文通过模拟研究了基于统计量F*，G*和J*的经验功效。

Summary statistics play an important role in point processes. They can be used to discriminate the types of different point patterns. Chen (2003) proposed three new functions F*, G* and J*; proved that G* and J* are very useful in studying large sample point patterns and can be used as summary statistics. But the powers of tests based on F*, G* and J* are unknown. In this simulation study we investigate empirically how the powers of tests based on F* compares to that of tests based on J* and G*.

1. Baddeley, A. J.(1999).Spatial sampling and censoring.Stochastic geometry: likelihood and computation,London:
2. Baddeley, A. J.(1999).A crash course in stochastic geometry.Stochastic geometry: likelihood and computation,London:
3. Baddeley, A. J.,Gill, R. D.(1993).Kaplan-Meier estimators for interpoint distance distribution of spatial point processes.
4. Chen, Jianbao(2003).Curtion University.
5. Doguwa, S. I.(1992).On the estimation of the point-object nearest neighbour distribution F(t) for point processes.J. Statist. Comp. Simul.,41,95-107.
6. Doguwa, S. I.,Upton, G. J. G.(1990).On the estimation of the nearest neighbour distribution, G(t), for point processes.Biometrical Journal,32,863-876.
7. Roder, W.(1975).A procedure for assessing point patterns without reference to area or density.The Professional Geographer,27,432-440.
8. Stoyan, D.,Kendall, W. S.,Mecke, J.(1995).Stochastic geometry and its Applications.Chichester:John Wiley and Sons.
9. Thönnes, E.,Van Lieshout, M. N. M.(1999).A comparative Study on the Power of Van Lieshout and Baddeley's J-function.Biometrical Journal,41,724-734.