Covsim: Simulating non-normal data with given covariance matrix
VITA | Bivariate case | Trivariate case | SEM example in 20 dimensions | Ordinal-categorical data | We assume that the underlying correlation in acontinuous bivariate distribution with standard normal marginals is$\rho=0.5$, and we discretize into three categories using thresholds$\tau_1=0$ and $\tau_2=1$. This means that we consider simulated dataof the form[X_i =\left{ \begin{matrix}1, & \text{if } \xi_i \leq \tau_1 \2, & \text{if } \tau_1 < \xi_i \leq \tau_2 \3, & \text{if } \xi_i > \tau_2 \\end{matrix}\right. | IG and Piecewise linear methods | IG | PLSIM | References
Rendered fromvignette.Rmdusingknitr::rmarkdown