摘要
研究了广义空间模型中单个异常值检验问题.分别在均值漂移模型和方差加权模型下导出了检验统计量的具体形式,并给出了在两种异常值模型下检验统计量的近似分布.最后,通过哥伦布市社区犯罪数据证明了方法的有效性.
This paper studies the score test for single outlier in general spatial model. Score test statistics are derived under the mean-shift model and the variance-weight model respectively. The approximate distributions of score statistics are obtained when outlier presents in two outlier models mentioned above. Finally, we illustrate our method using a real data of columbus neighborhood crime.
出处
《系统科学与数学》
CSCD
北大核心
2014年第1期106-118,共13页
Journal of Systems Science and Mathematical Sciences
基金
国家自然科学基金(11161053)资助课题
