Event element extraction method, computer equipment and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and provides an event element extraction method, computer equipment and a storage medium. After an event element entity is defined, a word vector and a vocabulary vector corresponding to the event element entity are extracted from target event content and are fused to obtain a first fusion vector, extracting a sentence vector, fusing the sentence vector, with the first fusion vector to obtain a second fusion vector, and finally performing prediction based on the second fusion vector by using a bidirectional long-short memory model to obtain an attribute value of the event element entity, thereby completing the extraction of the event element. As the first fusion vector combines the word vector and the vocabulary vector, the feature expression ability of the word vector is enhanced, extraction of the attribute value of the event element entity is facilitated, the second fusion vector fuses the sentence vector on the basis of the first fusion vector, the feature expression ability of the word vector is further enhanced, extraction of the attribute value of the event element entity is further improved, and the extraction accuracy of the event elements is high.

Description

technical field [0001] The invention relates to the technical field of information extraction, in particular to an event element extraction method, computer equipment and storage media. Background technique [0002] The Internet is the best source for collecting emergency news. After collecting the corresponding emergency news on the Internet, in order to have a clearer and more intuitive understanding of the emergency, it is necessary to use event extraction technology to automatically identify the event elements. [0003] The inventor found in the process of realizing the present invention that the existing event extraction technology generally adopts open source NER (Named Entity Recognition) tool, and the named entity recognition tool can directly identify the event time, location, event trigger words and event participants. class event elements, because these four types of event elements can be directly extracted from structured text, but named entity recognition tools ...

AI Technical Summary

Problems solved by technology

[0003] The inventor found in the process of realizing the present invention that the existing event extraction technology generally adopts open source NER (Named Entity Recognition) tool, and the named entity recognition tool can directly identify the event time, location, event trigger words and event participants. class event elements, because these four types of event elements can be directly extracted from structured text, but named entity recognition tools perform poorly on unstructured text, and secondly, for other event elements, such as event name, casualty Due to the need for further processing, the number of people, the list of casualties, and the main process, etc., cannot be directly extracted using named entity recognition tools, resulting in incomplete and ineffective extraction of event elements.

Images