Time series data augmentation method of small sample based on optimized generative adversarial network

被引:3
作者
Liu, Dongsheng [1 ]
Wu, Yuting [1 ]
Hong, Deyan [1 ]
Wang, Siting [2 ]
机构
[1] Zhejiang Gongshang Univ, Sch Comp & Informat Engn, Hangzhou, Peoples R China
[2] Zhejiang Gongshang Univ, Sch Management Engn & E Commerce, Hangzhou, Peoples R China
关键词
data augmentation; deep learning; GAN; particle swarm optimization; time series; STRATEGY;
D O I
10.1002/cpe.7331
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
In the time series classification task, for the problem that small samples have difficulty in capturing sequence features cause deep learning models have low classification accuracy. This article proposes a time series data augmentation method based on gated recurrent unit (GRU)-convolutional neural network (CNN) structure of generative adversarial network, denoted as GC-GAN. Specifically, the structure of GRU network in series with CNN is used as the generator of GAN model, while the discriminator part uses two-layer convolution network. In order to avoid the disappearance of gradient, Wasserstein distance is used to measure the distance between the generated distribution and the true distribution, and penalties are introduced in the loss function. Finally, the particle swarm algorithm is used to optimize the hidden variable input of the generator to approximate the real sequence to the greatest extent and obtain the best sample. The experimental results on the UCR time series dataset show that the long short-term memory network classification accuracy has been effectively improved after using the GC-GAN model, with a maximum improvement of 4.4%, indicating that the method proposed in this article has better data augmentation capabilities.
引用
收藏
页数:12
相关论文
共 45 条
[1]  
[Anonymous], 2016, ECMLPKDD WORKSHOP AD
[2]  
[Anonymous], Improved Training of Wasserstein GANs Montreal Institute for Learning Algorithms
[3]  
Arjovsky M, 2017, PR MACH LEARN RES, V70
[4]   Improving the accuracy of global forecasting models using time series data augmentation [J].
Bandara, Kasun ;
Hewamalage, Hansika ;
Liu, Yuan-Hao ;
Kang, Yanfei ;
Bergmeir, Christoph .
PATTERN RECOGNITION, 2021, 120
[5]   An under-sampled software defect prediction method based on hybrid multi-objective cuckoo search [J].
Cai, Xingjuan ;
Niu, Yun ;
Geng, Shaojin ;
Zhang, Jiangjiang ;
Cui, Zhihua ;
Li, Jianwei ;
Chen, Jinjun .
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2020, 32 (05)
[6]  
Chawla, DEEP LEARNING ANOMAL
[7]  
Cho K., 2014, ARXIV14061078, DOI [10.48550/arXiv.1406.1078, DOI 10.3115/V1/D14-1179]
[8]   A New Subspace Clustering Strategy for AI-Based Data Analysis in IoT System [J].
Cui, Zhihua ;
Jing, Xuechun ;
Zhao, Peng ;
Zhang, Wensheng ;
Chen, Jinjun .
IEEE INTERNET OF THINGS JOURNAL, 2021, 8 (16) :12540-12549
[9]   Personalized Recommendation System Based on Collaborative Filtering for IoT Scenarios [J].
Cui, Zhihua ;
Xu, Xianghua ;
Xue, Fei ;
Cai, Xingjuan ;
Cao, Yang ;
Zhang, Wensheng ;
Chen, Jinjun .
IEEE TRANSACTIONS ON SERVICES COMPUTING, 2020, 13 (04) :685-695
[10]   A Hybrid BlockChain-Based Identity Authentication Scheme for Multi-WSN [J].
Cui, Zhihua ;
Xue, Fei ;
Zhang, Shiqiang ;
Cai, Xingjuan ;
Cao, Yang ;
Zhang, Wensheng ;
Chen, Jinjun .
IEEE TRANSACTIONS ON SERVICES COMPUTING, 2020, 13 (02) :241-251