Multivariate LSTM-FCNs for time series classification

2018/01/31 by Fazle Karim, Somshubra Majumdar, Houshang Darabi +1 · 1,108 citations
Computer Science · Engineering · Mathematics · #Advanced Chemical Sensor Technologies #Anomaly Detection Techniques and Applications #Artificial intelligence #Block (permutation group theory) #Computer science #Convolutional neural network #Machine learning #Mathematics #Multivariate statistics #Pattern recognition (psychology) #Preprocessor #Series (stratigraphy) #Time Series Analysis and Forecasting #Time series #Univariate #cs.LG #stat.ML

paper · pdf · doi:10.1016/j.neunet.2019.04.014

published in Neural Networks 116, 237-245 (Elsevier BV) · 18 pages, 2 figures, 5 tables

openalex publication_date 2019/05/03 · arxiv created 2019/07/01 · arxiv updated 2019/07/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05

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.

Citations

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

Cited by

• A Hybrid Deep Learning based Carbon Price Forecasting Framework with Structural Breakpoints Detection and Signal Denoising
• Robotic Classification of Divers' Swimming States using Visual Pose Keypoints as IMUs
• Multi-scale Frequency-Aware Adversarial Network for Parkinson's Disease Assessment Using Wearable Sensors
• Training-Free Time Series Classification via In-Context Reasoning with LLM Agents
• Cognitive State Inference from VR Motion via Motion Foundation Model
• Deep Learning Inductive Biases for fMRI Time Series Classification during Resting-state and Movie-watching
• ReLATE+: Unified Framework for Adversarial Attack Detection, Classification, and Resilient Model Selection in Time-Series Classification
• Low-Cost Sensing and Classification for Early Stress and Disease Detection in Avocado Plants
• Enhancing Spatiotemporal Networks with xLSTM: A Scalar LSTM Approach for Cellular Traffic Forecasting
• A Contrastive Diffusion-based Network (CDNet) for Time Series Classification
• Highly adaptable deep-learning platform for automated detection and analysis of vesicle exocytosis
• Temporal cross-validation impacts multivariate time series subsequence anomaly detection evaluation
• Season-Independent PV Disaggregation Using Multi-Scale Net Load Temporal Feature Extraction and Weather Factor Fusion
• Wavelet Probabilistic Recurrent Convolutional Network for Multivariate Time Series Classification
• FIC-TSC: Learning Time Series Classification with Fisher Information Constraint
• TimeSAE: Causal Sparse Decoding for Faithful Explanations of Black-Box Time Series Models
• PHEATPRUNER: Interpretable Data-centric Feature Selection for Multivariate Time Series Classification through Persistent Homology
• Dynamic Graph-Like Learning with Contrastive Clustering on Temporally-Factored Ship Motion Data for Imbalanced Sea State Estimation in Autonomous Vessel
• Approaches and Applications of Early Classification of Time Series: A Review
• Inline quality assurance of hairpin welding using a machine learning approach based on photodiode signals and images
• XEM: An explainable-by-design ensemble method for multivariate time series classification
• ALT: A Python Package for Lightweight Feature Representation in Time Series Classification
• Towards Simple Machine Learning Baselines for GNSS RFI Detection

Related

• Multivariate Time Series Classification: A Deep Learning Approach
• Time Elastic Neural Networks
• Classification of multivariate weakly-labelled time-series with attention
• Multivariate Time Series Classification using Dilated Convolutional Neural Network
• FormerTime: Hierarchical Multi-Scale Representations for Multivariate Time Series Classification