Insights Into LSTM Fully Convolutional Networks for Time Series Classification
Insights Into LSTM Fully Convolutional Networks for Time Series Classification
复制标题
DOI：
10.1109/access.2019.2916828
复制
发表时间：
2019-01-01
期刊：
影响因子：
3.9
通讯作者：
Darabi, Houshang
中科院分区：
文献类型：
--
作者：
Karim, Fazle;Majumdar, Somshubra;Darabi, Houshang
长短期记忆全卷积神经网络(LSTM-FCNs)和注意力LSTM-FCN(ALSTM-FCN)在旧的加州大学河滨分校(UCR)时间序列库上的时间序列分类任务中取得了最先进的性能。然而，关于LSTM-FCN和ALSTM-FCN表现良好的原因还没有研究。在本文中，我们对LSTM-FCN和ALSTM-FCN进行了一系列烧蚀测试(3627个实验)，以更好地了解该模型及其各个子模块。对ALSTM-FCN和LSTM-FCN的烧蚀试验结果表明，LSTM和FCN块联用时的烧蚀效果更好。两种z归一化技术--单独对每个样本进行z归一化和对整个数据集进行z归一化--使用Wilcoxson符号等级检验进行了比较，结果显示在性能上存在统计学差异。此外，我们通过与LSTM-FCN在不应用维度洗牌时的性能比较，了解了影响维度洗牌对LSTM-FCN的影响。最后，我们展示了当LSTM块被门控递归单元(GRU)、基本神经网络(RNN)和稠密块取代时，LSTM-FCN的性能。
Long short-term memory fully convolutional neural networks (LSTM-FCNs) and Attention LSTM-FCN (ALSTM-FCN) have shown to achieve the state-of-the-art performance on the task of classifying time series signals on the old University of California-Riverside (UCR) time series repository. However, there has been no study on why LSTM-FCN and ALSTM-FCN perform well. In this paper, we perform a series of ablation tests (3627 experiments) on the LSTM-FCN and ALSTM-FCN to provide a better understanding of the model and each of its sub-modules. The results from the ablation tests on the ALSTM-FCN and LSTM-FCN show that the LSTM and the FCN blocks perform better when applied in a conjoined manner. Two z-normalizing techniques, z-normalizing each sample independently and z-normalizing the whole dataset, are compared using a Wilcoxson signed-rank test to show a statistical difference in performance. In addition, we provide an understanding of the impact dimension shuffle that has on LSTM-FCN by comparing its performance with LSTM-FCN when no dimension shuffle is applied. Finally, we demonstrate the performance of the LSTM-FCN when the LSTM block is replaced by a gated recurrent unit (GRU), basic neural network (RNN), and dense block.