机器学习方法在中医学传承研究中的应用及思考

Thinking on Application of Machine Learning Method in Inheritance of Traditional Chinese Medicine

推动中医药传承是保障中医学经久不衰的动力,随着中医学理论体系的不断丰富,传统的传承模式已经无法满足当今中医学传承的需求。在计算机技术飞速发展的当下,机器学习在中医学传承中凸显出了重要地位。梳理并总结比较基于机器学习的两种基本形式,即监督学习和无监督学习在相关中医药研究中的应用,认为监督学习更适合“病-症-证”的研究,在临床常见疾病的辨证模型、风险预测模型与证素危险因素研究中更具应用价值,而无监督学习更适合“方-药”研究和隐性知识的发现。基于对当前机器学习平台的比较,认为自研平台应当拓展算法类型,以实现与专业数据挖掘平台的更好结合。基于中医药传承研究与大数据科学结合的发展现状,认为机器学习将在中医理论的传承创新、临床决策和新药研发等方面产出更有意义的成果,但还需进一步拓展学科交叉的广度和深度,开发算法更加优化的机器学习平台。

The inheritance of theory and experience is the enduring power of traditional Chinese medicine(TCM).However,the traditional inheritance model can no longer keep pace with the enrichment of the theoretical system of TCM.With the rapid development of computer technology,machine learning has opened up a new way for the inheritance of TCM in this age.This paper explored the application of two basis forms of machine learning,that is supervised and unsupervised learning,in TCM researches.The results showed that supervised learning is more suitable for“disease-symptom-syndrome”researches concerning syndrome differentiation model,risk prediction model and risk factors of syndrome elements of common clinical diseases,whereas unsupervised learning is more valuable to the“formula-herb”researches and the reveal of tacit knowledge.After comparing between the current machine learning platforms,it is suggested that the self-developed platforms should enrich their algorithms in order to integrate with professional data mining platforms further.Based on the current inheritance status of TCM and its integration with big data science,it is believed that machine learning is a promising approach in the inheritance and innovation of TCM theories,the clinical decision makings and new drug researches.And in future,it is suggested to enlarge interdisciplinarity from both breadth and depth,and to develop more algorithms to perfect the machine learning platforms.