Mutual Information Mining for Component Law and Development of New Recipes of Topical Herbs for Atopic Dermatitis
With the development of industry, atopic dermatitis (AD) has becoming more and more prevalent in the pediatric population, with rates reportedly as high as 10̄20% in the England, and the 0.7% in the population from 6 to 20 years old in China. Modern medicine treatment for this disease includes corticosteroids, moisturizers, emollients, antihistamines, antibiotics, immunomodulators, immunosuppressants and etc. Recently, Complementary and Alternative Medicine, Includes Traditional Chinese medicine, has started to become a new research focus, and these studies mainly focused on reducing the amount of the use of corticosteroids and reducing the side effects of corticosteroids in patients of AD. Chinese medicine treatment of atopic dermatitis has a long history. However, so far, most of the Chinese medicine researches of atopic dermatitis focus on the summary of the doctor̉s experience. Code Shoppy The use of modern data mining methods to explore the topical treatment Recipes about atopic dermatitis of Traditional Chinese medicine topical herbs study have not yet to see. The aim of the study is to use the unsupervised data mining methods to analyze Traditional Chinese medicine topical herbs for treating atopic dermatitis, so as to provide references for clinical treatment of AD.
A.Source of Recipes A computer-based online search of relevant Chinese articles in the China Medical Code(Zhong-Hua-Yi-Dian), which is a datebase that concluded the Chinese ancient literature. The keywords of search term con-luded “Four bend wind(Si-Wan-Feng)”, “Milk ringworm(Nai-xuan)”, “Fetal sores(Tai-Lian-Chuang)”, “eyebrow sores(Lian-Mei-Chuang)” ,“Fetal ringworm(Tai-xuan)”, “Steeped sores(Jin-Yin-Chuang)”, “Blood winds sores(Xue-Feng-Chuang)”. B.Screening of Recipes We will exclude the recipes which contain only one Chinese topical herb, and selects recipes which contains two or more Chinese topical herbs. What’s more, the recipes which contain the oral plus topical herbs to treat AD are also excluded objects, because our aim is to explore the topical herbs on the impact of AD this time. C.Construction of Database We will take the Microsoft Net framework 3.5 as the development platform, and the computer program, all the Chinese herbs names in the recipes will become bi-value variable. Point-striking the Chinese herbs involved in a certain recipe by the mouse, that is, their values will be defined as “1” and the values of other drugs will be automatically expressed as “0”. From 84 Chinese articles, 84 Chinese drug recipes for the treatment of this class of diseases will be collected, 138 commonly-used Chinese drugs and a corresponding database of Chinese medicine topical herbs in recipes for AD will be constructed. D.Analysis Methods The data will be analyzed through the modified mutual information methods developed by this research group, for example, complex system entropy cluster method, and unsupervised hierarchical clustering method, and realizing quantitative description of drug correlation coefficient, and extraction of core combinations, and discovering new recipes. New recipes refer to those achieved according to the data mining methods, and these dates are not existed in previous database. For the mining methods, on the one hand, the core combinations of Chinese herbs for AD will be acquired by complex system entropy cluster, on the other hand, the core combinations attained will be deeply mined by unsupervised hierarchical clustering, which generate new recipes. The principles of the complex system entropy cluster and the hierarchical clustering are briefly introduced as follows. xThe complex system entropy cluster.