output_models/gaussian should either be updated to handle multidimensional Gaussians, or we build two separate output_models for 1D and nD.
Both estimation and generation methods should be fairly straightforward. Means are computed the same way, but just using vector additions. Covariance matrices are estimated the usual way with an outer product of the mean-free square vectors. Generation can be done using the numpy or scipy modules for multidimensional random variable generation.
output_models/gaussian should either be updated to handle multidimensional Gaussians, or we build two separate output_models for 1D and nD.
Both estimation and generation methods should be fairly straightforward. Means are computed the same way, but just using vector additions. Covariance matrices are estimated the usual way with an outer product of the mean-free square vectors. Generation can be done using the numpy or scipy modules for multidimensional random variable generation.