一种采用SOM架构的信息过滤系统语义扩展方法研究

Research of Semantic Extended Means of Information Filtering System Based on SOM Structure

论述一种基于自组地图(SOM:Self-Organizing Maps)架构的信息过滤系统语义扩展方法,利用系统本体对海量文本信息系统特征词之间的关联关系进行建模,计算文本信息的特征向量作为自组地图的输入,为用户实现一种直观的信息导航视图。相应实验结果表明,利用本体系统改善特征词权重的计算有效提高文本信息特征向量的分类和标示精度,减少系统的语义重复性,同时自组地图为用户提供的结果信息可以帮助其更为直观精确的定位到需求信息。

The article introduces a semantic extended means to realize an information filtering system based on SOM(Self-Organizing Maps) structure. Its work principle is to build a model of relation of feature terms in massive text information using system’s ontology. Then we generate the feature vector of text information as the input of SOM. This can take users a directive information navigating view. The result of experiment indicates that this means increases the classification and remark precision of text inform...