Sequential Pattern Recognition: Naive Bayes Versus Fuzzy Relation Method
Marek W. Kurzynski, Andrzej Żołnierek · 2006
In this paper two possibilities of taking into account the dependencies in the sequential pattern recognition task are considered. The first method is naive Bayes attempt adopted to the probabilistic model of sequential decision problem in which, the assumption of Markov dependence in the sequence of recognized patterns is made. The second one is the fuzzy relation approach, in which we omitted such not necessary correct assumptions. Furthermore, both methods were applied to the medical diagnostic task and the results of computer investigations are discussed