A Supervised Learning Framework for Prediction of Incompatible Herb Pair in Traditional Chinese Medicine.

Jiajing Zhu,Yongguo Liu,Shangming Yang,Shuangqing Zhai,Zhang Yi,Chuanbiao Wen

DOI: https://doi.org/10.1145/3269206.3269241

2018-01-01

Abstract:Adverse drug-drug interaction has been a critical issue for the development of drugs. In Traditional Chinese Medicine, adverse herb-herb interaction is a negative reaction in patients after the absorption of decoction of Incompatible Herb Pair (IHP). Recently, many methods have been proposed for IHP research, but most of them focused on revealing and analyzing the adverse reaction of some known IHPs, despite that a number of new IHPs have been discovered by accidents. Up to now, IHPs have been a serious threat to public health in the TCM medication. In this paper, we propose a novel supervised learning framework for potential IHP prediction. In this framework, we model the prediction task as a non-negative matrix tri-factorization problem, in which two important herb attributes (efficacy and flavor) and their correlation are incorporated to characterize the incompatible relationship among herbs. A hypothetical test method is adopted to evaluate the statistical significance of dissimilar characteristics of two attributes and the results are used as a regularization term to improve the accuracy of IHP prediction. Experiments on the real-world IHP dataset demonstrate that the proposed framework is very effective for prediction of potential IHPs.