网络科技信息中的知识对象行为识别方法

Method of Knowledge Object Behavior Identify in the Scientific and Technical Information of Network

在网络科技信息的情报价值判断中,知识对象行为具有重要的情报指示作用。在对语义图结构分析的基础上,针对不同的句法结构,设计了知识对象行为的识别规则以及逐层匹配的方法;对各层匹配结果中的指代关系进行处理,利用对象中心词和位置特征,替换匹配结果中的单个词,实现对象行为的识别;并对算法识别效果进行评测和分析,讨论其中存在的问题及未来的工作方向。

In the judgement of the intelligence value of scientific and technical information of network, knowledge object behavior plays an important role of intelligence indicating. Based on the analysis of semantic graph structure, this paper designs identification rules and layer by layer matching method of knowledge object behavior by aiming at different sytax structure. Based on the processing of coreference relation of the matching result in each layer, the paper replaces the single word of matching result by using the object headword and location characteristic, which in order to realize the identification of object behavior. The paper evaluates and analyzes the identification effect of algorithm, and discusses the existing problems and working direction in the future.