Person Event Detection Method in Computer Discipline Domain Based on BiGRU and CNN in Series

Abstract

The knowledge graph of computer discipline domain plays a critical role in computer education, and the person event is an important part of the discipline knowledge graph. Adding person events to the graph will make the discipline knowledge graph richer and more interesting, and enhance enthusiasm of students for learning. The most crucial step in building the person event knowledge graph is the extraction of trigger words. Therefore, this paper proposes a method based on the serial fusion of gated recurrent neural network and convolutional neural network (SC-BiGRU-CNN) for person event detection in the computer discipline domain. We extract the global features of the text from the person event sentences through the BiGRU model, and input the extracted global features into the CNN model to further extract the fine-grained features of the text. And then the extracted features are used to classify the event trigger words. In addition, a dataset (CD-PED) for person event detection in the computer discipline domain is constructed to obtain trigger words and their types. We perform experiments on the public dataset MAVEN and the domain dataset CD-PED, respectively. The experimental results show that our approach has significantly improved the [Formula: see text] value compared with the baseline model on the domain dataset CD-PED.