Many studies have applied association rule algorithms in mining formulae in traditional Chinese medicine (TCM) so far. However, in these studies, the treatment effect of prescriptions is not considered. In the meantime, a proper minimum support threshold is needed for these association rule algorithms, which is difficult for doctors. This paper proposes TCMM (traditional Chinese medicine miner), a novel algorithm for mining the top-K weighted association rule which does not require the users to set a support threshold in advance. In addition, TCMM can analyze TCM formulae with treatment effect. After conducting several experiments on TCM formulae for diabetic nephropathy, we verify that our algorithm can facilitate doctors to conduct new drug discovery and make better clinical medicine selection.
展开▼
