randomizeX {MMLR}

Transformation of regressors' matrix X. Data preparation stage for simulation

Description

Additional function to be used for simulation purposes (academical or research). Transforming the matrix of regressors according to user preferences. Random disturbances (according to a chosen distribution) are entered in the initial values of the matrix X. The expectation of the resulting matrix coincides with the initial matrix X.

Usage

randomizeX(X, p = 1, k0 = 1, k1 = 0.5, k2 = 1, k3 = 1)

Arguments

X

Matrix (n x k), n - number of observations, k - number of columns (k - 1 - number of regressors). Note, that 1st column contains only ones (1) (intercept)

p

Scalar (from 1 to +inf), random number for simulation. The default value is 1

k0

Scalar. Number from 1 to 3 (distribution selection). k0 = 1 - uniform distribution (RV Z ~ U (K1, k2)). k0 = 2 - exponential distribution (RV Z ~ exp(lambda)). k0 = 3 - Gamma distribution (RV Z ~ gamma(shape, rate)). The default value is k0 = 1.

k1

Scalar. 1) If k0 = 1, then k1 is a left boundary of uniform distribution (RV Z ~ U(k1, k2)). 2) If k0 = 2, then k1 is a parameter lambda of exponential distribution (RV Z ~ exp(lambda)). 3) If k0 = 3, then k1 is a shape parameter of Gamma distribution (RV Z ~ gamma(shape, rate)). The default is k1 = 0.5.

k2

Scalar. 1) If k0 = 1, then k2 is a right boundary of uniform distribution (RV Z ~ U(k1, k2). 2) If k0 = 3, then k2 is a rate parameter of Gamma distribution (RV Z ~ gamma(shape, rate)). The default is k2 = 1.

k3

Scalar. The number of digits after the comma when rounded. The default value is 1.

Details

Random perturbations are added to the initial values of matrix X elements ($X_i,j$ + rnd), which are distributed according to a chosen distribution (possible alternatives: uniform, exponential and gamma distribution).

Value

New transformed matrix of regressors (n x k), according to user preferences.

Examples

Xtest <- cbind(rep_len(1,10),c(2,5,7,3,1,1,2,2,3,6), c(5,9,1,2,3,2,3,5,2,2))
randomizeX(Xtest,1,1,1,2,2)