领域本体信息源选取方法研究与实现

Research on method of information sources selection for domain ontology building and its implementation

构造领域本体所需的信息源选取方法的研究对解决本体的构造质量、构造效率等问题,以及推广与发展领域本体有着重要意义.传统的信息源文档选取方法只考虑概念因素,不能很好地解决该问题.因此,首先利用抽象方法分析了领域本体所需信息源具有的概念性、关系性和预测性等特点.然后,针对这些特点分别采用改进的V SM方法、基于本体关系距离的方法以及神经网络方法计算文档权值.最后,通过编写的软件O nM aker产生模拟数据得到概念、关系和预测3个权值,从而计算出每个文档权值,并使用与"湿地保护"相关的真实文档验证该模型,达到了较好排序选取的效果.

Method of information source selection is important to build domain ontology with regard to improving the ontology quality and efficiency, and develop the ontology. The classical methods only take concepts into account, and fall short of solving practical problems well. Therefore, an abstract method analysis is firstly used to analyze the characteristics of information sources, such as conceptuality, relativity and predictability, and then considering these properties, three methods - the improved vector space model （VSM）, ontology relation distance and neural network, are introduced to calculate these characteristics weights, respectively. Finally, the simulated data is generated by implementing software OnMaker, and three weights of concept, relation and prediction are obtained, and in the following every document weight is calculated. Combined with a real document data set of ＂Wetland Protection＂, the model is tested and a good order effect on the document selection is attained.