作者: Gregor Chliamovitch , Alexandre Dupuis , Bastien Chopard
DOI: 10.3390/E17063738
关键词: Principle of maximum entropy 、 Maximum-entropy Markov model 、 Maximum entropy thermodynamics 、 Mathematical optimization 、 Maximum entropy spectral estimation 、 Maximum entropy probability distribution 、 Markov chain 、 Rényi entropy 、 Mathematics 、 Statistical physics 、 Entropy rate
摘要: We develop ideas proposed by Van der Straeten to extend maximum entropy principles Markov chains. focus in particular on the convergence of such estimates order explain how our approach makes possible estimation transition probabilities when only short samples are available, which opens way applications non-stationary processes. The current work complements an earlier communication providing numerical details, as well a full derivation multi-constraint two-state and three-state matrices.