PRTools Contents

PRTools User Guide

gentrunk

GENTRUNK

Generation of Trunk's classification problem of 2 Gaussian classes

    A = GENTRUNK(N,K)

Input
 N Dataset size, or 2-element array of class sizes (default: [50 50]).
 K Dimensionality of the dataset to be generated (default: 10).

Output
 A Dataset.

Description

Generation of a K-dimensional 2-class dataset A of N objects. Both classes  are Gaussian distributed with the idenity matrix as covariance matrix.  The means of the first class are defined by ua(j) = 1/sqrt(j). The means  for the second class are ub = -ua. These means are such that the Nearest  Mean Classifier always shows peaking for a finite training set.

Reference(s)

1. G.V. Trunk, A Problem of Dimensionality: A Simple Example, IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 1, pp. 306-307, 1979
2. A.K. Jain, R.P.W. Duin, and J. Mao, Statistical Pattern Recognition: A Review, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, pp. 4-37, 2000.

Example(s)

 a = gentrunk([1000 1000],200);
 e = clevalf(a,nmc,[1:9 10:5:25 50:25:200],[5 5],25);
 plote(e)

See also

datasets, prdatasets,

PRTools Contents

PRTools User Guide

This file has been automatically generated. If badly readable, use the help-command in Matlab.