A Novel Method for Signal Sequence Classification Based on Markov Reward Models

The efficiency and accuracy of signal sequence classification have always been the ultimate goals of researchers. However, it is difficult for existing methods to meet both requirements at the same time. This paper proposes a new signal sequence classification method based on Markov Reward Model (MRM) to solve the above problem. Firstly, a deterministic probabilistic finite automaton, learned from training sequence dataset, is transformed into a discrete time Markov Chain; then, leveraging JS divergence, a MRM is constructed; and finally, sequence classification is achieved on MRM efficiently and accurately. This method can be applied to many practical signal processing applications.