Chinese Emergency Event Recognition Using Conv-RDBiGRU Model.

Haoran Yin,Jinxuan Cao,Guodong Wang,Luzhe Cao

doi:10.1155/2020/7090918

Haoran Yin, Jinxuan Cao + Show 2 more

Open Access

https://doi.org/10.1155/2020/7090918

Copy DOI

Abstract

In view of the weak generalization of traditional event recognition methods, the limitation of dependence on field knowledge of expert, the longer train time of deep neural network, and the problem of gradient dispersion, the neural network joint model, Conv-RDBiGRU, integrated residual structure was proposed. Firstly, text corpus is preprocessed by word segmentation and stop words processing and uses word embedding to form the matrix of word vectors. Then, local semantic features are extracted through convolution operation, and deep context semantic features are extracted through RDBiGRU. Finally, the learned features are activated by softmax function and the recognition results are output. The novelty of work is that we integrate residual structure into recurrent neural network and combine these methods and field of application. The simulation results show that this method improves precision and recall of Chinese emergency event recognition, and the F-value is better than other methods.

Highlights

As a manifestation of information, an event is defined as the objective fact that specific people and objects interact with each other at a specific time and place [1]. e Internet is full of all kinds of disorderly emergency event news, which is intermingled with other news, and these other news will hinder clear cognition of users about emergency event and relevant researchers’ work in classification and storage [2], so how to realize emergency event recognition in the network is one of the problems that need to be solved at present
In order to validate the effectiveness of the Conv-RDBiGRU model proposed in the emergency event recognition and the recognition result is superior to the method of other model, multiple comparative experiments are performed in this paper, including hyperparameters adjustment to optimize model, comparison with different emergency event recognition models in other papers, and test in different datasets
In the task of Chinese emergency event recognition, ConvRDBiGRU neural network model is proposed in this paper

Summary

Introduction

As a manifestation of information, an event is defined as the objective fact that specific people and objects interact with each other at a specific time and place [1]. e Internet is full of all kinds of disorderly emergency event news, which is intermingled with other news, and these other news will hinder clear cognition of users about emergency event and relevant researchers’ work in classification and storage [2], so how to realize emergency event recognition in the network is one of the problems that need to be solved at present. With the development of deep learning technology, in the field of emergency event recognition, artificial neural network has obtained more and more attention and has been successfully applied in this field to solve practical problems. In [13], the classifier which combines support vector machine with radial basis function neural network to improve the reliability of an event recognition result is proposed. E model employs a Bidirectional Long Short Term Memory (BLSTM) neural network and multilevel attention mechanism for event extraction and achieves good results in biomedical event recognition [18]. Is method can introduce more reference information to the output data, and, at the same time, make up the influence of vanishing gradient along with the network deepened, so as to improve Chinese emergency event recognition effect In the model, extracting local information feature of text through convolution operation, capturing deep context feature information of text corpus using DBiGRU part, and using the residual structure to change the framework structure that data source for this network layer are only from previous layer in the traditional plain network. e changed framework structure introduces a design that is similar to “shortcut” type [20]. e source data input skips multiple hidden layers and is added directly to the data output section. is method can introduce more reference information to the output data, and, at the same time, make up the influence of vanishing gradient along with the network deepened, so as to improve Chinese emergency event recognition effect

Related Work

Findings

Experiments and Discussion

Conclusion and Future Work