基于Hadoop的中医症状群分类应用

CLASSIFICATION OF TRADITIONAL CHINESE MEDICINE SYNDROME BASED ON HADOOP PLATFORM

传统单一的KNN算法来挖掘中医病案症状-证型规律,存在占用内存空间大、建模时间长、复杂度高、大数据量无法处理等问题。提出基于Hadoop平台的并行化计算中医症状群分类方法。在Hadoop分布式计算平台中,利用MapReduce计算框架,并行化实现KNN分类算法。实验结果表明,基于Hadoop的中医症状群预测分类效率更高,能更有效地指导临床实践。

When the traditional single KNN algorithm is used to analyse the syndrome pattern of TCM（ Traditional Chinese Medicine）,a series of problem that large space being occupied,long modelling time,high calculating complexity and failing to deal with large data arises, therefore the parallel computing method of TCM symptom classification based on Hadoop platform was proposed. In the Hadoop distributed computing platform, the KNN classification algorithm was implemented in parallel with the MapReduce computing framework. The experimental results show that the prediction and classification of TCM symptoms based on Hadoop is more efficient and it can guide clinical practice more effectively.