Multivariate LSTM-FCNs for Time Series Classification

Abstract

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.

arXiv comment18 pages, 2 figures, 5 tables

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Neural Machine Translation by Jointly Learning to Align and Translate
• Squeeze-and-Excitation Networks
• LSTM Fully Convolutional Networks for Time Series Classification
• How to Construct Deep Recurrent Neural Networks
• Multivariate Time Series Classification with WEASEL+MUSE
• Ultra-Fast Shapelets for Time Series Classification
• Autoregressive Kernels For Time Series

Cited by in corpus (58)

• Benchmarking Multivariate Time Series Classification Algorithms
• Action Transformer: A Self-Attention Model for Short-Time Pose-Based Human Action Recognition
• Improving Position Encoding of Transformers for Multivariate Time Series Classification
• Insights into LSTM Fully Convolutional Networks for Time Series Classification
• XCM: An Explainable Convolutional Neural Network for Multivariate Time Series Classification
• Neural network based country wise risk prediction of COVID-19
• Hamiltonian neural networks for solving equations of motion
• Multivariate Time Series Classification with WEASEL+MUSE
• Approaches and Applications of Early Classification of Time Series: A Review
• TodyNet: Temporal Dynamic Graph Neural Network for Multivariate Time Series Classification
• Digital Twin: Values, Challenges and Enablers
• Intra-domain and cross-domain transfer learning for time series data -- How transferable are the features?
• Time-Series Classification in Smart Manufacturing Systems: An Experimental Evaluation of State-of-the-Art Machine Learning Algorithms
• XEM: An Explainable-by-Design Ensemble Method for Multivariate Time Series Classification
• Unsupervised Recurrent Federated Learning for Edge Popularity Prediction in Privacy-Preserving Mobile Edge Computing Networks
• Benchmarking Online Sequence-to-Sequence and Character-based Handwriting Recognition from IMU-Enhanced Pens
• Forecasting the abnormal events at well drilling with machine learning
• A Shapelet-based Framework for Unsupervised Multivariate Time Series Representation Learning
• Deep Efficient Continuous Manifold Learning for Time Series Modeling
• Adjusting for Autocorrelated Errors in Neural Networks for Time Series
• Transient Classifiers for Fink: Benchmarks for LSST
• Evaluation of (Un-)Supervised Machine Learning Methods for GNSS Interference Classification with Real-World Data Discrepancies
• A Time-Temperature Dataset for the Strawberry Cold Chain Across Multiple Shipments and Locations
• MedLens: Improve Mortality Prediction Via Medical Signs Selecting and Regression
• A Generalised Signature Method for Multivariate Time Series Feature Extraction
• Collaborative Attention Mechanism for Multi-View Action Recognition
• Residual Attention Net for Superior Cross-Domain Time Sequence Modeling
• The Deep Arbitrary Polynomial Chaos Neural Network or how Deep Artificial Neural Networks could benefit from Data-Driven Homogeneous Chaos Theory
• I see artifacts: ICA-based EEG artifact removal does not improve deep network decoding across three BCI tasks
• Intelligently Augmented Contrastive Tensor Factorization: Empowering Multi-dimensional Time Series Classification in Low-Data Environments
• Fusing Structure from Motion and Simulation-Augmented Pose Regression from Optical Flow for Challenging Indoor Environments
• Learning a Proposal Classifier for Multiple Object Tracking
• Spikebench: An open benchmark for spike train time-series classification
• IEEE BigData 2021 Cup: Soft Sensing at Scale
• Multi-task CNN Behavioral Embedding Model For Transaction Fraud Detection
• Condition Assessment of Stay Cables through Enhanced Time Series Classification Using a Deep Learning Approach
• Fast CRDNN: Towards on Site Training of Mobile Construction Machines
• RTFN: A Robust Temporal Feature Network for Time Series Classification
• Diagnostic Spatio-temporal Transformer with Faithful Encoding
• Adversarial Attacks on Multivariate Time Series
• Correlative Channel-Aware Fusion for Multi-View Time Series Classification
• Seq2Tens: An Efficient Representation of Sequences by Low-Rank Tensor Projections
• A Hybrid Deep Learning Model for Predictive Flood Warning and Situation Awareness using Channel Network Sensors Data
• LLT: An R package for Linear Law-based Feature Space Transformation
• Season-Independent PV Disaggregation Using Multi-Scale Net Load Temporal Feature Extraction and Weather Factor Fusion
• Prior knowledge distillation based on financial time series
• Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings
• ML-Based Approach for NFL Defensive Pass Interference Prediction Using GPS Tracking Data
• On Technical Trading and Social Media Indicators in Cryptocurrencies' Price Classification Through Deep Learning
• RED CoMETS: An ensemble classifier for symbolically represented multivariate time series
• Adaptive Law-Based Transformation (ALT): A Lightweight Feature Representation for Time Series Classification
• RTFN: Robust Temporal Feature Network
• Instance Explainable Temporal Network For Multivariate Timeseries
• Effective Human Activity Recognition Based on Small Datasets
• SummerTime: Variable-length Time SeriesSummarization with Applications to PhysicalActivity Analysis
• iWash: A Smartwatch Handwashing Quality Assessment and Reminder System with Real-time Feedback in the Context of Infectious Disease
• Mimic: An adaptive algorithm for multivariate time series classification
• ODE guided Neural Data Augmentation Techniques for Time Series Data and its Benefits on Robustness