CorrelatedDesignMatrix

Function for use in simulations. Created design matrix of categorical variables with correlated columns

Computes solutions for linear and logistic regression models with potentially high-dimensional categorical predictors. This is done by applying a nonconvex penalty (SCOPE) and computing solutions in an efficient path-wise fashion. The scaling of the solution paths is selected automatically. Includes functionality for selecting tuning parameter lambda by k-fold cross-validation and early termination based on information criteria. Solutions are computed by cyclical block-coordinate descent, iterating an innovative dynamic programming algorithm to compute exact solutions for each block.

Daniel Grose

CatReg

Solution Paths for Linear and Logistic Regression Models with
Categorical Predictors, with SCOPE Penalty

Benjamin Stokell

Rajen Shah

CorrelatedDesignMatrix function

<dl><dt>n</dt>
<dd>Number of observations</dd>
<dt>cov_mat</dt>
<dd>p x p covariance matrix. Controls correlations of pairs of marginally U[0,1] random variables that are subsequently binned to assign categories for each variable</dd>
<dt>c</dt>
<dd>Number of categories within each variable</dd></dl>

Arguments

Create a design matrix of categorical variables with correlated columns. — CorrelatedDesignMatrix

<dl>

<dt>n</dt>
<dd>Number of observations</dd>


<dt>cov_mat</dt>
<dd>p x p covariance matrix. Controls correlations of pairs of marginally U[0,1] random variables that are subsequently binned to assign categories for each variable</dd>


<dt>c</dt>
<dd>Number of categories within each variable</dd>

</dl>

CorrelatedDesignMatrix: Create a design matrix of categorical variables with correlated columns.

Description

Usage

Value

Arguments

Examples