基于领域本体的股票资讯新闻自动语义标注技术

An automatic semantic-taggingtechnique on stock information news based on domain ontologies

目的利用领域本体来提高股票资讯新闻自动标注结果的质量。方法在构建股票资讯领域本体的基础上,首先使用最大正向匹配技术来提取股票资讯新闻中的关键信息,然后将关键信息与股票资讯领域本体进行匹配,最后利用领域本体标注出与该股票资讯新闻相关的股票。结果比较实验结果表明基于领域本体的股票资讯新闻标注技术的准确率明显高于基于关键字的标注技术。结论将股票资讯领域本体中的知识用于标注股票资讯新闻可以确定更多同股票资讯新闻相关的潜在股票,从而提高标注结果的质量。

Purposes-To improve the automatic annotating quality of stock news through the domain ontologies.Methods-On the basis of the domain ontologies on stock information,the maximum forward matching technique is first used to extract key information from stock news,and then,the key information is matched with its domain ontologies.Finally,the domain ontologies are employed to mark the stocks related to the stock information news.Result-The experimental results show that the accuracy of the proposal is significantly higher than that of the keyword-based annotating technique.Conclusion-Using the knowledge in the stock domain ontology for annotating the stock news can identify more related stocks,which improves the quality of the tagging results.