一种用于文本分类的逻辑推理方法探讨  

Research on A Logical Inference Method for Text Classification

作　　者：李东艳[1] 张永奎[1] 

机构地区：[1]山西大学,太原030006

出　　处：《电脑开发与应用》2004年第4期7-9,共3页Computer Development & Applications

基　　金：山西省归国人员基金 (2 0 0 2 0 0 4 )资助

摘　　要：针对使用布尔逻辑推理的 OCAT方法进行文本分类时存在的不确定文本比重较大的问题 ,提出带权重的 WOCAT文本分类算法。算法将文本属性权重信息用于推理及判别过程 ,在原有判别效果基础上对不确定文档进行进一步判别。A weighted OCAT method of text classification is presented for the problem that the results of the OCAT method that uses Boolean logical inference have many undecided documents.This method adopts the weight information of text attribute to the process of inference and decision,and gives a further identification to the undecided documents on the basis of the original decision result.The analyzed results show that the ratio of correct classification in this approach is higher than that of the OCAT method.

关 键 词：人工智能 文本分类 逻辑推理 OCAT 权值 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]