Multivariate LSTM-FCNs for Time Series Classification
This work addresses classification tasks like activity recognition for applications needing efficient deployment on memory-constrained systems, but it is incremental as it builds on existing univariate models.
The authors tackled multivariate time series classification by adapting univariate LSTM-FCN models with a squeeze-and-excitation block, resulting in improved accuracy that outperforms most state-of-the-art models with minimal preprocessing.
Over the past decade, multivariate time series classification has received great attention. We propose transforming the existing univariate time series classification models, the Long Short Term Memory Fully Convolutional Network (LSTM-FCN) and Attention LSTM-FCN (ALSTM-FCN), into a multivariate time series classification model by augmenting the fully convolutional block with a squeeze-and-excitation block to further improve accuracy. Our proposed models outperform most state-of-the-art models while requiring minimum preprocessing. The proposed models work efficiently on various complex multivariate time series classification tasks such as activity recognition or action recognition. Furthermore, the proposed models are highly efficient at test time and small enough to deploy on memory constrained systems.