GENDATV

Generation of a very large dataset

    [A,U] = GENDATV(M,C,K,R,T)
    A = GENDATV(M,U)

Input
M	Number of objects to be generated (def. 10^4)
C	Number of classes (def. 10^2)
K	Number of features (def 10)
R	std. dev of class distributions (def 1)
S	Std. dev. of class priors (def 0)
U	Dataset with class means, class priors and R

Output
A	C-class dataset with M objects and K features
U	Dataset with class means, class priors and R

Description

This routine generates an arbitrarily large dataset of C classes by first  generating the class means from a standard K-dimensional normal  distribution. The class priors are genetated from a N(1,S) distribution,  After that their sum is normalized to one. The classes themselves have  spherical normal distributions with an equal standard deviation R for all  features and all classes.

In U the class means, the class priors and R are returned, such that  similar data from the same distribution can be generated.

See also

datasets, prdatasets,