THREE LAYER MODEL FOR AN ORIENTAL MEDICAL DIAGNOSIS BASED ON BYY BINARY INDEPENDENT FACTOR ANALYSIS

摘要

The three layer model for an oriental medical diagnosis via BYY Binary Independent Factor Analysis. In this model, both unsupervised learning and supervised learning method are used in the steps. Specially BYY harmony learning and model selection are provided for Binary Independent Analysis on networks with one hidden layer binary units. We develop an automatic oriental medical diagnostic system based on BYY Binary Independent Analysis. Moreover, experiments are made on the application to the diagnostic problem of 'liver disease' with the medical sampling data in the oriental medical diagnostic perspectives.