multivariate-normal

Multivariate Normal

SimpleMVN1

Simulate and fit MVN data with a compound symmetric correlation structure.

See YouTube video 🎥 link in the in the main overview.

SimpleMVN2

Simulate and fit MVN data with an unstructured correlation structure. Demonstrates how to keep the variance-covariance matrix positive definite by parameterizing it using the special TMB functionUNSTRUCTURED_CORR.

The two examples introduce:

• namespace density to enable estimation of multivariate normal distribution parameters
• .rows() to extract the number of rows of a matrix
• .cols() to extract the number of columns of a matrix
• .row(i) to extract the ith row of a matrix
• << to assign values to elements of a matrix
• typecasting to use a row of a matrix as a vector
• log transformation to keep a parameter positive
• shifted logistic transformation to keep a parameter between -1 and 1