| Traditional Chinese medicine (TCM) has a long history and has been accepted as the unique medical method of China. TCM is also a branch of clinical medicine.With the experience accumulated from clinic, it forms theoretical knowledge and uses it back into clinic, which is the fundamental pattern of its academic development. The prescription and its proportion rule formed in the long-term clinic is the basic knowledge of aquiring effective prescription and diagnosis-treatment techniques. Therefore, it's favorable for the treatment of disease and the development of TCM to find the common and effective prescription from the knowledge accumulated. With so large amounts of clinical data, including the prescriptions and the proportion rules, we hope to investigate some effective and important herb-symptom subsets by data mining.In this paper, we propose a method called BTMDR that can find subsets with high herb-symptom relativity from a large number of prescriptions and then analyze these subsets to see which of these subsets is truly effective.The method consists of two core modules, including Bilcustering algorithm and MDR. The main research contents of this paper are as follows.(1) By the studty of biclustering, we use cHawk algorithm to investigate the subset of effective and important herb-symptom interactions. Experimental evaluation confirms that this approach is able to effectively get the sub-module of herb-symptom.(2) Studying of the MDR algorithm, we use MDR software to analyze the result of Biclustering algorithm by cross-validation and investigate effective and important herb-symptom modules.(3) Combinating Biclustering algorithm with MDR algorithm ,we have implemented BTMDR algorithm. It's a good method which can analyze and investigate effective and important herb-symptom modules. We test and get clinical interpretation by using the method in the insomnia and infertility databases. Experimental evaluation confirms we can get meaningful results.
|
PDF Full Text Request