基于蒙特卡洛算法的中医药特性研究及应用

Research and Application of Chinese Medicine Characteristics Based on Monte Carlo Algorithm

用数据挖掘的方法分析了445味中草药的信息,发现其内容有缺失部分为功能主治、性味、采制等。借助蒙特卡洛算法,利用概率统计理论可以得到其余中药材性味与功能主治、生长地、植物形态的关联以及概率,通过比较相关已知内容,得出缺失的信息。由于信息量大,要先对信息特征进行数字化处理,用聚类分析的方法研究中药分布信息,通过使用SPSS软件中的聚类方法,比较中药各属性间的相似度,聚类得出最适合的分类情况,之后通过聚类谱系图以及最终聚类中心表分析出分布情况。最后,找出主治呼吸道感染的止咳中药,及患者所处地点、发病季节以及并发症状,筛选出符合治疗条件的中药,并根据药性及相关文献参考配制成止咳方案。

By using the method of data mining, 445 flavors of herbal medicine information were analyzed,and found missing parts of the contents were indications, taste, collect and process and etc. By means of Monte Carlo algorithm, using the theory of probability and statistics, the rest of herbal’s taste and function, growth, plant morphology and association probability could be obtained. By comparing the relevant known content, to obtain the missing information. Because of the amount of information, the first was to digitize the information characteristics, analyzed Chinese medicine distribution information with clustering analysis. By using clustering method in SPSS software, compared similarity calculation between each attribute, obtained the most suitable clustering classification,then analyzed the distribution through cluster analysis and clustering center table. Finally, found the main treatment of respiratory tract infection of Chinese medicine cough, and the location of the patients, the onset of the season and complications, screening out the Chinese medicine eligible for treatment, according to the nature and the related literature reference to prepare the antitussive regimen.