A stochastic learning based approach for automatic medical diagnosis using HMM toolbox in scilab environment

In this work, automatic medical diagnosis system of sleep apnea syndrome is presented. This system is based on hidden Markov models (HMMs) Scilab toolbox. Conventional as well as a new simulated annealing based approaches to train HMMs are incorporated. The inference method of this system translates event state value into common interpretation as a pathophysiological state. The interpretation is extended to sequences of states in time to obtain a pathophysiological state-space trajectory. Some of the measurements of the respiratory activity issued by the technique of polysomnography are considered for offline or online detection of different sleep apnea syndromes. Experimental results using respiratory clinical data and some future perspectives of our work are presented