Citation: Chen, Z.; Huang, K.; Wu, L.;
Zhong, Z.; Jiao, Z. Relational Graph
Convolutional Network for
Text-Mining-Based Accident Causal
Classification. Appl. Sci. 2022, 12,
2482. https://doi.org/10.3390/
app12052482
Academic Editors: Nikos D. Lagaros
and Vagelis Plevris
Received: 30 January 2022
Accepted: 24 February 2022
Published: 27 February 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Relational Graph Convolutional Network for
Text-Mining-Based Accident Causal Classification
Zaili Chen
1,2,†
, Kai Huang
2,3,†
, Li Wu
1,
*, Zhenyu Zhong
1
and Zeyu Jiao
2,
*
1
Faculty of Engineering, China University of Geosciences, Wuhan 430074, China; zl.chen@cug.edu.cn (Z.C.);
zy.zhong@giim.ac.cn (Z.Z.)
2
Guangdong Key Laboratory of Modern Control Technology, Institute of Intelligent Manufacturing,
Guangdong Academy of Sciences, Guangzhou 510070, China; kirehuang@gmail.com
3
School of Economics and Management, Beihang University, Beijing 100191, China
* Correspondence: lwu@cug.edu.cn (L.W.); zy.jiao@giim.ac.cn (Z.J.)
† These authors contributed equally to this work.
Abstract:
Accident investigation reports are text documents that systematically review and analyze
the cause and process of accidents after accidents have occurred and have been widely used in the
fields such as transportation, construction and aerospace. With the aid of accident investigation
reports, the cause of the accident can be clearly identified, which provides an important basis for
accident prevention and reliability assessment. However, since accident record reports are mostly
composed of unstructured data such as text, the analysis of accident causes inevitably relies on a lot
of expert experience and statistical analyses also require a lot of manual classification. Although, in
recent years, with the development of natural language processing technology, there have been many
efforts to automatically analyze and classify text. However, the existing methods either rely on large
corpus and data preprocessing methods, which are cumbersome, or extract text information based
on bidirectional encoder representation from transformers (BERT), but the computational cost is
extremely high. These shortcomings make it still a great challenge to automatically analyze accident
investigation reports and extract the information therein. To address the aforementioned problems,
this study proposes a text-mining-based accident causal classification method based on a relational
graph convolutional network (R-GCN) and pre-trained BERT. On the one hand, the proposed method
avoids preprocessing such as stop word removal and word segmentation, which not only preserves
the information of accident investigation reports to the greatest extent, but also avoids tedious
operations. On the other hand, with the help of R-GCN to process the semantic features obtained by
BERT representation, the dependence of BERT retraining on computing resources can be avoided.
Keywords: accident causal classification; accident investigation reports; text mining; R-GCN; BERT
1. Introduction
Accident investigation reports are usually text documents formed by professional
investigators or teams through visits, conversations, viewing video surveillance and an-
alyzing recorded data after accidents occur [
1
] and have been widely used in aviation,
construction, transportation and other fields [
2
]. The process and consequences of the
accident recorded in the reports can be leveraged by experts to analyze the cause of the acci-
dent, which is of great significance for preventing the recurrence of the accident or forming
the accident response plan [
3
]. However, the current analysis of accident investigation
reports mainly relies on expert experience to manually determine the cause of the accident,
which requires a lot of work, and the accuracy is affected by the subjective experience of
experts [
4
]. On 29 October 2018, an Indonesian Lion Air Boeing 737 MAX8 plane carrying
189 passengers and crew was flying from Jakarta’s Soekarno Hatta International Airport
to Penang Port, Bangka Belitung Province. The plane lost contact 13 min after takeoff and
was later confirmed to have crashed in the waters off Karawang, West Java province [
5
].
Appl. Sci. 2022, 12, 2482. https://doi.org/10.3390/app12052482 https://www.mdpi.com/journal/applsci
