基于关联规则和遗传算法Web文档分类.pdfVIP

_______________________________________________________________________________ 1 Classify Web Document by Genetic Algorithm with Association Rules Tang Changj ie(唐常杰) Zhang Tianqing, （张天庆）, Hu Rong （胡蓉）, Yuan Chang-an （元昌安）,Chen Anlong （陈安龙） (Compu ter Depar tment, Sichuan University, Chengdu 610064, China) chj tang@ Abstract: Classifying Web Document such as BBS, HTML and e-mail, etc., is an important task for web application. To solve this problem, this paper presents following results: ( 1) Proposes a new text classification method called Classification by Genetic Algorithm with Association Rules Method (CGAA method). (2) Other than previous work, the fitness function are applied under the guidance of the association rules mined by Apriori_CGAA algorithm. (3) Realizing a family of genetic procedures such as CGAA _Roulette_Selection, CGAA_Xover and CGAA _binaryMutation and giving extensive experiments with real data. (4)The experiment show that the CGAA algorithm is superior to other common methods. A Best-Vector with a score 3513.6 can be achieved after running CGAA algorithm after 50 generations. Key words: Chinese document classification, Genetic Algorithm, Association rules, Natural language processing, CGAA method (Classification by Genetic Algorithm with Association Rules) 基于关联规则和遗传算法的Web 文档分类 唐常杰 张天庆,胡蓉,元昌安,陈安龙 四川大学计算机学院 摘要 Web 文档分类，例如BBS、HTML、e-mail 的分类是Web 引用中的重要任务。为了解决这一问题， 该翁做了下列工作：（1）提出了一个用于中文文本自动分类的、称为基于关联规则和遗传算法的Web 文档 分类的计算方法（Classification by Genetic Algorithm with Association Rules Method 缩写为CGAA ）。（2 ）不 同于前人的路线，适应度函数的在关联规则制导下工作，而