作者: Wilbert C. M. Kallenberg , Teresa Ledwina
DOI: 10.1080/01621459.1999.10473844
关键词: Statistical hypothesis testing 、 Model selection 、 Score test 、 Mathematics 、 Statistics 、 Correlation 、 Linear correlation 、 Monte Carlo method 、 Data-driven 、 Copula (probability theory)
摘要: We introduce new rank tests for testing independence. The procedures are sensitive not only grade linear correlation, but also correlations of higher-order polynomials. number polynomials involved is determined by the data. Model selection combined with application score test in selected model. Whereas well-known as Spearman's or Hoeffding's may completely break down alternatives that dependent have low greater power stability. Monte Carlo results clearly show this behavior. Theoretical support obtained proving consistency tests.