A `MultivariateLogNormalPopulationModel` is a mutlidimensional lognormal distribution that allows for nonzero covariance between the dimensions. There are different options to implement this distribution, but for inference it appears to be the most efficient to implement it using a Cholesky decomposition of the covariance matrix, see https://www.google.com/search?q=multivariate+normal+distribution+cholesky&oq=multivariate+normal+distribution+ch&gs_lcrp=EgZjaHJvbWUqCAgDEAAYFhgeMgYIABBFGDkyCAgBEAAYFhgeMggIAhAAGBYYHjIICAMQABgWGB4yCAgEEAAYFhgeMggIBRAAGBYYHjIICAYQABgWGB4yCAgHEAAYFhgeMg0ICBAAGIYDGIAEGIoFMg0ICRAAGIYDGIAEGIoF0gEJMTAzMzNqMGo3qAIAsAIA&sourceid=chrome&ie=UTF-8
A
MultivariateLogNormalPopulationModelis a mutlidimensional lognormal distribution that allows for nonzero covariance between the dimensions.There are different options to implement this distribution, but for inference it appears to be the most efficient to implement it using a Cholesky decomposition of the covariance matrix, see https://www.google.com/search?q=multivariate+normal+distribution+cholesky&oq=multivariate+normal+distribution+ch&gs_lcrp=EgZjaHJvbWUqCAgDEAAYFhgeMgYIABBFGDkyCAgBEAAYFhgeMggIAhAAGBYYHjIICAMQABgWGB4yCAgEEAAYFhgeMggIBRAAGBYYHjIICAYQABgWGB4yCAgHEAAYFhgeMg0ICBAAGIYDGIAEGIoFMg0ICRAAGIYDGIAEGIoF0gEJMTAzMzNqMGo3qAIAsAIA&sourceid=chrome&ie=UTF-8