A method to discover compatibility regulation of Chinese herbs based on the combination of clustering analysis and herbs nature

Online Herbs Shopping Project

A method to discover compatibility regulation of Chinese herbs based on the combination of clustering analysis and herbs nature

the purpose of this paper is to find a data mining method to clustering herbs. The results of data mining are to confirm the relation among herbs in a Chinese medicine prescription. Firstly, the similarity between herbs is measured by the attributes of herbs without the direction based on the prior knowledge. Secondly, the herbs are classified by K­Means algorithm. Lastly, the relations are confirmed by the combination of the classification and the Chinese medicine theory. The experiment illustrates that the method is effective, practical and generalized.

https://codeshoppy.com/shop/product/online-herbs-shopping/

Online Herbs Shopping Project

. Chinese herb data model 1) Before using data mining method, relational data model is established according to database management system about Prescriptions for Universal Relief. Chinese herbs in a prescription are clustered to analyze with herb nature including of five flavors, four properties and meridian entry. They are in the next part concretely. 2) Five flavors are sour, bitter, sweet, hot and salty. 3) Herb properties are cold, little cold, cool, little cool, medium, little warm, warm and hot. 4) Meridian entries are liver, heart, spleen, lung, kidney, pericardium, gallbladder, small intestine, stomach, large intestine, bladder and three cavities. B. Data pretreatment Chinese herb data is processed with digitization before cluster. The herb nature is assigned from cold to hot in turn. Cold is -1, little cold is -0.75, cool is -0.5, little cool is -0.25, medium is 0, little wann is 0.25, wann is 0.5 and hot is 1. The flavors of herb include sour, bitter, sweet, hot and salty. In accordance with the rule of promoting and constraining mutually to concept of the five elements, the membership function value of certain taste is 1, little taste is 0.25. If an herb has two tastes, the later is 0.5. If two herbs are generation, the second one is 0.5. If two herbs are antagonism, the second one is -0.5. There are 12 meridial entries. If one herb has certain meridial entry, the membership function value is 1.

All the Chinese herbs in a formula Are divided into three classifications when K is equal to three. The results of cluster analysis on relieving cough pill are in TableIV. From the cluster analysis results, the first classification herbs property is warm not cold. The second classification herbs property is cold and their flavor is bitter. The 3rd classification herbs flavor is sour and their channel tropism are lung and kidney. The results are in conformity with Chinese medicine theory. And the results show that this sort standard is considered from multiple attributes of herbs.

First classified data is gained by cluster analysis, and then monarch herb, minister herb, assistant herb and guide herb are judged by heuristic rule, at last the herb properties are analyzed by Chinese medicine theory and the results are verified further more. Eight prescriptions such as cassia twig Decoction , Minor Bluegreen dragon decoction, Minor Bupleurum Decoction, Regullating the Middle Pill, Four Gentlemen Decoction, Pink Decoction and Activating Blood Decoction 231 for Reviving Health are selected to be analyzed by k=2 cluster analysis. The results of cluster analysis on eight kinds of Chinese medicine formulae is in Table V and VI. The decision rule is obtained by Table V and VI and the Chinese medicine theory: I) If one Chinese herb is in the same classification, it would be thought as monarch herb. For example that cassia twig Decoction and Regullating the Middle Pill are. 2) If two Chinese herbs are in one classification, they would be monarch herb and minister herb, such as Minor Bupleurum decoction, Minor Bupleurum decoction, Four Gentlemen decoction, Pink decoction Minor Bluegreen dragon decoction and so on Read more

Categories