贝叶斯网络在中医药领域的研究现状及展望  

Research Status and Prospects of Bayesian Networks in Field of Traditional Chinese Medicine

作　　者：卢恩仕 韩明光 刘祖发[1] LU Enshi;HAN Mingguang;LIU Zufa(Wangjing Hospital Affiliated of China Academy of Chinese Medical Sciences,Beijing China 100102;School of Mathematical Sciences of Peking University,Beijing China 100871)

机构地区：[1]中国中医科学院望京医院,北京100102 [2]北京大学数学科学学院,北京100871

出　　处：《中医学报》2022年第2期316-320,共5页Acta Chinese Medicine

基　　金：北京中医药科技发展资金项目(JJ2018-93);中国中医科学院“优势病种-医院制剂-新药”研发专项项目(ZZ15-XY-CT-08);中国中医科学院望京医院科研课题项目(WJYY2016-PY-01)。

摘　　要：贝叶斯网络技术的信息处理过程与中医辨证思维相吻合,目前已广泛应用于中医证候规律探究、名老中医经验传承、中药药性研究、疾病干预措施疗效评价及现代化中医诊断技术中,将复杂多样的中医数据转变成客观、量化的标准。目前中医辅助诊疗系统已建立并不断完善,促进了临床辅助诊疗及远程医疗的发展。贝叶斯网络技术在中医药研究的诸多领域发挥着重要作用,但仍存在有待改进之处:首先,样本数据的支持力度不足,扩大样本数量、提高样本质量是进一步实现数据分析客观化的必要选择;其次,贝叶斯网络算法利用频率进行,一旦出现频率较低的症状、证候要素及证型时,算法会在变量筛选的过程中将其舍弃,从而导致信息获取不全面,影响结果的准确性,故需进一步获取更多的高质量样本,加强对精进算法的探索。随着研究的不断深入,今后亟需建立智能化的中医药诊疗系统,提供个体化的治疗处方及调理建议,以发挥中医药"未病先防,既病防变"的诊疗优势。The information processing process of Bayesian network technology is consistent with TCM dialectical thinking.It has been widely used in the exploration of TCM syndrome rules, the inheritance of the experience of famous TCM doctors, the research on the properties of Chinese medicines, the evaluation of the efficacy of disease intervention measures, and the modern TCM diagnosis technology, and transform complex and diverse Chinese medicine data into objective and quantitative standards.At present, the TCM auxiliary diagnosis and treatment system has been established and continuously improved, which has promoted the development of clinical auxiliary diagnosis and treatment and telemedicine.Bayesian network technology plays an important role in many fields of Chinese medicine research, but there are still shortcomings: First of all, the support of sample data is insufficient.Enlarging the number of samples and improving the quality of samples are necessary choices to further realize the objectification of data analysis.Secondly, the Bayesian network algorithm uses frequency.Once the symptoms, syndrome elements, and syndrome types with lower frequency appear, the algorithm will discard them in the variable screening process, which will lead to incomplete information acquisition and affect the accuracy of the results.Therefore, it is necessary to further strengthen the exploration of refined algorithms.With the continuous deepening of research, there is an urgent need to establish an intelligent Chinese medicine diagnosis and treatment system in the future, to provide individualized treatment prescriptions and conditioning suggestions, so as to give full play to the diagnosis and treatment advantages of "preventing the disease before it occurs and preventing the progress of the existing diseases" in TCM.

关 键 词：中医药 贝叶斯网络 数据挖掘 现代中医诊断技术 辅助诊疗 

分 类 号：R2-03[医药卫生—中医学]