The application of subspace clustering to cooperation of chinese herbal medicines
Yufeng Xu, Zhou Li-hua, Qing Xiao, Chen JinSong · 2008
The cooperation of Chinese herbal medicines is not simple mixture of several remedies. The employ of remedies has certain correlation. However, the conventional clustering techniques are unlikely to reveal these relationships on the data of Chinese herbal medicines, because the data are always on the high dimensionality and spread around. Not all Chinese herbal medicines are associated with each other. The subspace clustering techniques not only can identify interesting clusters, but also find the attributes related to the cluster. In this paper, a subspace clustering algorithm HARP is used to analyze the employ of remedies on cooperation of Chinese rheum prescription. The interesting clusters have been discovered on various subspaces. Many potential laws on cooperation of Chinese prescription are revealed.