目的：使用概率潜在语义分析（PLSA）算法研究中药配伍方案，为中药处方发现提供新途径。方法：基于丰富的中医药文献数据，从临床治疗缺血性心脑血管疾病的方剂出发，使用PLSA算法，筛选出治疗缺血性心脑血管疾病的新药候选方，并从药理学的角度对候选方进行分析，最后结合专家智慧在候选方的基础上给出具有开发潜力的新药处方。结果：筛选出治疗该病的中药核心组合，并证明了PLSA算法用于中药处方发现具有一定的可靠性和稳定性。同时专家从计算机筛选出的处方中选出有开发前景的两首方。结论：PLSA算法在治疗缺血性心脑血管疾病中药处方发现中有着广阔的应用前景。

This study was aimed to find a new available method for Chinese medicine prescription research and development. Probability latent semantic analysis (PLSA) was used to calculate some new herbal combinations. First, 3547 TCM clinic papers about curing ischemic cardiovascular and cerebrovascular disease were selected from the Clinic Literature Database. Second, the name of syndrome, prescription, and herb were regulated according to the Traditional Chinese Medicine Subject Terms Table and the Chinese Medicine Standard Table. Third, PLSA was used to find out combinations of herbs which have the potential of the semantic relations. Finally, some interesting conclusions were obtained from PLSA calculating results through advices given by experts. It was concluded that PLSA calculation of Chinese medicine can provide a new method for Chinese medicine prescription research and development.

科学技术部国家科技重大新药创制专项（2009ZX09301原005-001）：中药新药发现、新药设计和信息平台，负责人：崔蒙；中国中医科学院基本科研业务费自主选题项目（ZZ050305）：基于支持向量机的中药组分药理作用预测模型的研究，负责人：雷蕾。