KISTCM: knowledge discovery system for traditional Chinese medicine

摘要: Objective: Traditional Chinese Medicine (TCM) provides an alternative method for achieving and maintaining good health. Due to the increasing prevalence of TCM large volume data accumulated though thousands years, there is urgent need efficiently effectively explore this information its hidden rules with knowledge discovery in database (KDD) techniques. This paper describes design development a system as well newly proposed KDD techniques integrated system. Methods: A novel Knowledge dIscovery System (KISTCM) developed by incorporating several mining techniques, primarily including medicine dependency relationship algorithm, efficacy dimension reduction algorithm based on neural networks, exploring relationships between formulae syndromes using gene expression programming (GEP), approach discovering properties terms nature, taste meridian herbal dosage employing effect degree function calculate each property. Results: Representative experimental cases are used evaluate performance. Encouraging results obtained, previously unknown designers experiment runners. Experiments demonstrate that KISTCM has powerful analysis capabilities, useful tool underlying formulae. Our successfully discover from data, which new direction discovery. From experts' perspective, accuracy improvement, these compare favorably other existing The could be expected practice TCM, e.g., assisting physicians prescribing or automatically distinguishing minister assistant herbs formula.