Proposition of common classifier construction for pattern recognition with context task

This paper deals with the concept of information fusion and its application to the contextual pattern recognition task. The concept of the recognition based on the probabilistic model are presented. The machine learning algorithm based on statistical tests for the recognition of controlled Markov chains is shown. Idea of information unification via transforming the expert rules into the learning set is derived. Some experimental results of obtained methods are shown and proposed concepts are applied to the real medical decision problem.