Simulation Model

Three sets of variables are calculated: explanatory, intermediate and response variables.

Usage

sim1(N, tau=0.8, prob=0.5)

Arguments

N

number of simulated observations.

tau

describes the connection between explanatory and intermediate variable.

prob

starting value, a probability for a Bernoulli experiment.

Details

Explanatory, intermediate and response variables are bivariate variables. Let one explanatory variable be denoted by x and one intermediate variable be denoted by w respectively. 10 explanatory and 10 intermediate variables are simulated, which define one response variable z. The probability tau describes the connection between one explanatory and one intermediate variable: tau := P(w=0|x=1) = P(w=1|x=0).

Value

A list of three sets of variables

X

N*10 matrix of 10 explanatory variables.

Y

N*10 matrix of 10 intermediate variables.

zr

response vector.

References

David J. Hand, Hua Gui Li, Niall M. Adams (2001), Supervised classification with structured class definitions, Computational Statistics and Data Analysis, 36, 209-225

Author

Andrea Peters <Peters.Andrea@imbe.imed.uni-erlangen.de>

Note

The simulation model is appropriate for the indirect classification approach and described in Hand et al. (2001)