摘要
文章提出一种自动生成概念空间的方法。首先通过SOM神经网络,对文本进行聚类,之后从结果中提取反映各类文本内容的概念,用于标注文本的类别,再通过模糊聚类进行概念自动抽象与归纳形成概念空间,用于文本的管理。SOM本身是无监督的学习方式,在设定好参数后,经过训练自动生成文本空间与概念空间的映射图。相关试验和结果表明概念空间对文本有很好的分类管理功能,便于文本检索。
A method of the space of the concept on text document automatically generated is presented.Firstly,this paper clusters the text document through SOM and gains the concept of text for marking the classification,then uses the fuzzy clustering to automatically generate and sum up the concept space for managing the text document.The result can be used to cluster the Chinese document and generates an index of the cluster automatically.SOM is an unsupervised -learning neural -network method that produces a mapping from text space into concept space.Some experiments and test results show that the space of concept does well in arranging the classification of the text and is convenient to information retrieval.
出处
《计算机工程与应用》
CSCD
北大核心
2002年第7期63-65,88,共4页
Computer Engineering and Applications
基金
国家自然科学基金资助(编号:90104021
60173017)
北京市自然科学基金资助(编号:4011003)
