Semi-supervised Optimal Recursive Filtering and Smoothing in Non-Gaussian Markov Switching Models

•Exact filtering and smoothing in a non-Gaussian Markov switching state-space model.•Non-Gaussianity is processed by using Copulas and non-Gaussian margins.•Identification of model parameters is made semi-supervised by using ICE principle.•Automatic selection of shape for copulas and margins by using appropriate metrics.