An Incident Identification Method
Based on Improved RCNN
Han He, Haijun Zhang
(&) , Sheng Lv, and Bingcai Chen
School of Computer Science and Technology, Xinjiang Normal University,
Urumqi 830054, China
zhjlp@163.com
Abstract. An emergency is a sudden and harmful event. It is of great significance to quickly identify the event and reduce the harm caused by the event. In
this paper, the current advanced recurrent convolutional neural networks
(RCNN) are utilized, but the traditional model cannot effectively identify the
event, and the accuracy rate is not good enough. In order to solve this problem,
the recurrent neural network and activation function part of the traditional model
are improved, and through experimental comparison, the optimal model in the
training model is selected. Finally, the accuracy of the model is 90%, the recall
rate is 92.55%, and the F1 value, a metric that combines accuracy and recall, is
91.26%, which proves that the improved model has good effects.
Keywords: RCNN Á LSTM Á Event recognition Á RELU Á Deep learning
1 Introduction
China, a populous country, has experienced frequent emergencies. Since 2009, China
has identified the detection of emergencies as one of the key research projects.
Emergency [1] is an event that has no symptoms but causes harm to society and people,
and it is divided into four types: natural disasters, accident disasters, public health
events, and social security events [2]. The location, time, type, etc. of an emergency are
often beyond people’s imagination and generally occur when people are not prepared.
In this case, it is precisely because of this urgency and uncertainty that China hopes to
identify emergencies as early as possible through relevant research and take measures
to minimize casualties.
In recent years, the technology of event recognition based on deep learning has
attracted the attention of many scholars. In this paper, the recurrent convolutional
neural network in deep learning is studied, and an improved model is proposed to solve
the non-convergence of the model training process and to increase the recognition
accuracy of the model.
The remainder of this paper is arranged as follows: the second section introduces
the related works about event recognition; the third section describes the principle of
the revised model on the traditional RCNN; the fourth part gives experiments and
discussions on the proposed method. In the end, fifth section presents the conclusion
and future works.
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 96–104, 2020.
https://doi.org/10.1007/978-981-15-0187-6_11
Based on Improved RCNN
Han He, Haijun Zhang
(&) , Sheng Lv, and Bingcai Chen
School of Computer Science and Technology, Xinjiang Normal University,
Urumqi 830054, China
zhjlp@163.com
Abstract. An emergency is a sudden and harmful event. It is of great significance to quickly identify the event and reduce the harm caused by the event. In
this paper, the current advanced recurrent convolutional neural networks
(RCNN) are utilized, but the traditional model cannot effectively identify the
event, and the accuracy rate is not good enough. In order to solve this problem,
the recurrent neural network and activation function part of the traditional model
are improved, and through experimental comparison, the optimal model in the
training model is selected. Finally, the accuracy of the model is 90%, the recall
rate is 92.55%, and the F1 value, a metric that combines accuracy and recall, is
91.26%, which proves that the improved model has good effects.
Keywords: RCNN Á LSTM Á Event recognition Á RELU Á Deep learning
1 Introduction
China, a populous country, has experienced frequent emergencies. Since 2009, China
has identified the detection of emergencies as one of the key research projects.
Emergency [1] is an event that has no symptoms but causes harm to society and people,
and it is divided into four types: natural disasters, accident disasters, public health
events, and social security events [2]. The location, time, type, etc. of an emergency are
often beyond people’s imagination and generally occur when people are not prepared.
In this case, it is precisely because of this urgency and uncertainty that China hopes to
identify emergencies as early as possible through relevant research and take measures
to minimize casualties.
In recent years, the technology of event recognition based on deep learning has
attracted the attention of many scholars. In this paper, the recurrent convolutional
neural network in deep learning is studied, and an improved model is proposed to solve
the non-convergence of the model training process and to increase the recognition
accuracy of the model.
The remainder of this paper is arranged as follows: the second section introduces
the related works about event recognition; the third section describes the principle of
the revised model on the traditional RCNN; the fourth part gives experiments and
discussions on the proposed method. In the end, fifth section presents the conclusion
and future works.
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 96–104, 2020.
https://doi.org/10.1007/978-981-15-0187-6_11