ppeda

Projection pursuit exploratory data analysis. This function implements projection pursuit exploratory data analysis using the Posse chi-square index and the moment index.

Input Arguments:

• Z is an n by p matrix of sphered observations..
• C is the size of the starting neighborhood for each search.
• HALF is the number of steps without an increase in the index, at which time the neighborhood is halved.
• M is the number of random starts.
• PPI specifies which index to use: 'chi' for the chi-square index and 'mom' for the moment indes.

Output Arguments:

• ALPHA and BETA are vectors that span the plane where structure was found using PPEDA.
• PPM is the value of the projection pursuit index for this plane.

Synopsis

[ALPHA,BETA,PPM] = PPEDA(Z,C,HALF,M,PPI)

References:

Christian Posse. 1995. 'Projection pursuit explortory data analysis,' Computational Statistics and Data Analysis, vol. 29, pp. 669-687.
Christian Posse. 1995. 'Tools for two-dimensional exploratory projection pursuit,' J. of Computational and Graphical Statistics, vol 4 pp. 83-100.

Jones, M. C. and R. Sibson. 1987. “What is projection pursuit” (with discussion), Journal of the Royal Statistical Society, Series A, 150:1–36.