Earticle

초록

영어

Predicting disaster classification on time is critical to mitigate the damage. Since identifying disaster types requires large amounts of data, and real-world data are often imbalanced, there are many recent works addressing data imbalance problems using generative models. However, if the process of generating text data based on disaster classes and severity is not handled improperly, the quality of the data can be degraded as well as the performance of classification predictions. In this paper, we propose a scheme for generating data with enhanced quality using text based on labels such as informational value of text and severity of disasters. Our experiment results verify the quality of data through the comparisons of prediction performance between various machine learning models.

목차

Abstract
I. INTRODUCTION
II. RELATED WORKS
A. Genrative Models
B. Disaster Classification
III. DISASTER CONDITIONAL VAE
A. DC-VAE Architecture
B. Evaluation of Data Quality
IV. EXPERIMENTS
A. CrisisMMD dataset
B. Hyperparameter setting
C. Performance Evaluation
D. Synthetic data evaluation
E. Result Analysis
V. CONCLUSION
ACKNOWLEDGMENT
REFERENCES

저자

• Hyunseok Jung [ Department of Artificial Intelligence Ajou University ]
• Jiheon Choi [ Department of Artificial Intelligence Ajou University ]
• Jongwon Park [ Department of Digital Media Ajou University ]
• Sehui Baek [ Department of e-Business Ajou University ]
• Sangyoon Oh [ Department of Artificial Intelligence Ajou University ] Corresponding Author

참고문헌