Temporal factor analysis (TFA): stable-identifiable family, orthogonal flow learning, and automated model selection

Temporal factor analysis has been elaborated in a broad scope of the state space model called the stable-identifiable TFA family, and a number of typical situations and the corresponding stability and identifiability have been discussed. Then, orthogonal flow algorithms are suggested for learning, with not only a good numerical property in computation but also an easy implementation of automatic model selection during learning