Generate correlated systems of statistical equations which represent repeated measurements or clustered data. These systems contain either: a) continuous normal, nonnormal, and mixture variables based on the techniques of Headrick and Beasley (2004) <DOI:10.1081/SAC120028431> or b) continuous (normal, nonnormal and mixture), ordinal, and count (regular or zeroinflated, Poisson and Negative Binomial) variables based on the hierarchical linear models (HLM) approach. Headrick and Beasley's method for continuous variables calculates the beta (slope) coefficients based on the target correlations between independent variables and between outcomes and independent variables. The package provides functions to calculate the expected correlations between outcomes, between outcomes and error terms, and between outcomes and independent variables, extending Headrick and Beasley's equations to include mixture variables. These theoretical values can be compared to the simulated correlations. The HLM approach requires specification of the beta coefficients, but permits group and subjectlevel independent variables, interactions among independent variables, and fixed and random effects, providing more flexibility in the system of equations. Both methods permit simulation of data sets that mimic realworld clinical or genetic data sets (i.e. plasmodes, as in Vaughan et al., 2009, <10.1016/j.csda.2008.02.032>). The techniques extend those found in the 'SimMultiCorrData' and 'SimCorrMix' packages. Standard normal variables with an imposed intermediate correlation matrix are transformed to generate the desired distributions. Continuous variables are simulated using either Fleishman's thirdorder (<DOI:10.1007/BF02293811>) or Headrick's fifthorder (<DOI:10.1016/S01679473(02)000725>) power method transformation (PMT). Simulation occurs at the componentlevel for continuous mixture distributions. These components are transformed into the desired mixture variables using random multinomial variables based on the mixing probabilities. The target correlation matrices are specified in terms of correlations with components of continuous mixture variables. Binary and ordinal variables are simulated by discretizing the normal variables at quantiles defined by the marginal distributions. Count variables are simulated using the inverse CDF method. There are two simulation pathways for the multivariable type systems which differ by intermediate correlations involving count variables. Correlation Method 1 adapts Yahav and Shmueli's 2012 method <DOI:10.1002/asmb.901> and performs best with large count variable means and positive correlations or small means and negative correlations. Correlation Method 2 adapts Barbiero and Ferrari's 2015 modification of the 'GenOrd' package <DOI:10.1002/asmb.2072> and performs best under the opposite scenarios. There are three methods available for correcting nonpositive definite correlation matrices. The optional error loop may be used to improve the accuracy of the final correlation matrices. The package also provides function to check parameter inputs and summarize the simulated systems of equations.
Package details 


Author  Allison Cynthia Fialkowski 
Maintainer  Allison Cynthia Fialkowski <[email protected]> 
License  GPL2 
Version  0.1.0 
URL  https://github.com/AFialkowski/SimRepeat 
Package repository  View on CRAN 
Installation 
Install the latest version of this package by entering the following in R:

Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.