Sun, J. Significance levels in exploratory projection pursuit. (English) Zbl 0753.62067 Biometrika 78, No. 4, 759-769 (1991). Summary: In exploratory data analysis, projection pursuit methods explore the ‘nonlinear’ structure of high dimensional data. It is useful to have a significance test to help us decide whether apparent structure is real or just the effect of noise. Monte Carlo methods can be helpful to achieve this, but, unfortunately, in this case they are computationally expensive.Under a suitable null hypothesis, we derive a theoretical approximation for the \(P\) value associated with Friedman’s projection pursuit index [J. H. Friedman, J. Am. Stat. Assoc. 82, 249-266 (1987; Zbl 0664.62060)]. The result of Monte Carlo simulations is compared with our analytical result. Some practical aspects are discussed. Cited in 1 ReviewCited in 13 Documents MSC: 62M40 Random fields; image analysis 62H99 Multivariate analysis Keywords:testing multivariate normality; Gaussian random fields; exploratory data analysis; projection pursuit methods; significance test; approximation; P value; Friedman’s projection pursuit index; Monte Carlo simulations Citations:Zbl 0664.62060 PDFBibTeX XMLCite \textit{J. Sun}, Biometrika 78, No. 4, 759--769 (1991; Zbl 0753.62067) Full Text: DOI