N-BEATS: Neural basis expansion analysis for interpretable time series forecasting

2019/05/24 by Boris N. Oreshkin, Oreshkin, Boris N., Dmitri Carpov +5 · 195 citations
Computer Science · Decision Sciences · Mathematics · #Forecasting Techniques and Applications #Stock Market Forecasting Methods #Time Series Analysis and Forecasting #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1905.10437

arxiv created 2020/02/20 · arxiv updated 2020/02/24

Abstract

We focus on solving the univariate times series point forecasting problem using deep learning. We propose a deep neural architecture based on backward and forward residual links and a very deep stack of fully-connected layers. The architecture has a number of desirable properties, being interpretable, applicable without modification to a wide array of target domains, and fast to train. We test the proposed architecture on several well-known datasets, including M3, M4 and TOURISM competition datasets containing time series from diverse domains. We demonstrate state-of-the-art performance for two configurations of N-BEATS for all the datasets, improving forecast accuracy by 11% over a statistical benchmark and by 3% over last year's winner of the M4 competition, a domain-adjusted hand-crafted hybrid between neural network and statistical time series models. The first configuration of our model does not employ any time-series-specific components and its performance on heterogeneous datasets strongly suggests that, contrarily to received wisdom, deep learning primitives such as residual blocks are by themselves sufficient to solve a wide range of forecasting problems. Finally, we demonstrate how the proposed architecture can be augmented to provide outputs that are interpretable without considerable loss in accuracy.

Citations

• Forecasting Sales by Exponentially Weighted Moving Averages

Cited by

• Chain-of-thought Reviewing and Correction for Time Series Question Answering
• SANet: A Semantic-aware Agentic AI Networking Framework for Cross-layer Optimization in 6G
• Expert System for Bitcoin Forecasting: Integrating Global Liquidity via TimeXer Transformers
• QuantFlow: A Federated Mamba-Based Post-Transformer Foundation Model for Time-Series Forecasting
• Differencing the Diffusion Trajectory toward Uncertain Components for Time Series Forecasting
• Deep Generative Spatiotemporal Engression for Probabilistic Forecasting of Epidemics
• Dual-Prototype Disentanglement: A Context-Aware Enhancement Framework for Time Series Forecasting
• DecoKAN: Interpretable Decomposition for Forecasting Cryptocurrency Market Dynamics
• Mitigating Exposure Bias in Risk-Aware Time Series Forecasting with Soft Tokens
• Adaptive Information Routing for Multimodal Time Series Forecasting
• FRWKV:Frequency-Domain Linear Attention for Long-Term Time Series Forecasting
• STELLA: Guiding Large Language Models for Time Series Forecasting with Semantic Abstractions
• A Multi-View Multi-Timescale Hypergraph-Empowered Spatiotemporal Framework for EV Charging Forecasting
• TS2Vec-Ensemble: An Enhanced Self-Supervised Framework for Time Series Forecasting
• Tiny-TSM: Efficiently Training a Lightweight SOTA Time Series Foundation Model
• SimDiff: Simpler Yet Better Diffusion Model for Time Series Point Forecasting
• Structured Noise Modeling for Enhanced Time-Series Forecasting
• PeriodNet: Boosting the Potential of Attention Mechanism for Time Series Forecasting
• Moirai 2.0: When Less Is More for Time Series Forecasting
• LAD-BNet: Lag-Aware Dual-Branch Networks for Real-Time Energy Forecasting on Edge Devices
• DeepBooTS: Dual-Stream Residual Boosting for Drift-Resilient Time-Series Forecasting
• Predicting the Future by Retrieving the Past
• Synapse: Adaptive Arbitration of Complementary Expertise in Time Series Foundational Models
• ForecastGAN: A Decomposition-Based Adversarial Framework for Multi-Horizon Time Series Forecasting
• Multi-period Learning for Financial Time Series Forecasting
• Efficient Linear Attention for Multivariate Time Series Modeling via Entropy Equality
• Hydra: Dual Exponentiated Memory for Multivariate Time Series Analysis
• Discovering EV Charging Site Archetypes Through Few Shot Forecasting: The First U.S.-Wide Study
• Contextual Deconvolution for Variance-Stable Demand Sensing: Kernel-Modulated Operators in Promotional Retail
• E-STGCN: extreme spatio-temporal graph convolutional networks for air quality forecasting
• ARIMA_PLUS: Large-scale, Accurate, Automatic and Interpretable In-Database Time Series Forecasting and Anomaly Detection in Google BigQuery
• OneCast: Structured Decomposition and Modular Generation for Cross-Domain Time Series Forecasting
• Spectral Temporal Graph Neural Network for Multivariate Time-series Forecasting
• NetBurst: Event-Centric Forecasting of Bursty, Intermittent Time Series
• Hierarchical Graph Networks for Accurate Weather Forecasting via Lightweight Training
• Chronos-2: From Univariate to Universal Forecasting
• Crisis-Resilient Portfolio Management via Graph-based Spatio-Temporal Learning
• InvDec: Inverted Decoder for Multivariate Time Series Forecasting with Separated Temporal and Variate Modeling
• SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series
• Unsupervised Anomaly Prediction with N-BEATS and Graph Neural Network in Multi-variate Semiconductor Process Time Series
• Abstain Mask Retain Core: Time Series Prediction by Adaptive Masking Loss with Representation Consistency
• QKCV Attention: Enhancing Time Series Forecasting with Static Categorical Embeddings for Both Lightweight and Pre-trained Foundation Models
• Forecasting with sktime: Designing sktime's New Forecasting API and Applying It to Replicate and Extend the M4 Study
• Online Time Series Forecasting with Theoretical Guarantees
• Xihe: Scalable Zero-Shot Time Series Learner Via Hierarchical Interleaved Block Attention
• Beyond Accuracy: Are Time Series Foundation Models Well-Calibrated?
• Tackling Time-Series Forecasting Generalization via Mitigating Concept Drift
• CoRA: Covariate-Aware Adaptation of Time Series Foundation Models
• Metropolis-Scale Road Network Datasets for Fine-Grained Urban Traffic Modeling
• CauchyNet: Compact and Data-Efficient Learning using Holomorphic Activation Functions
• STaTS: Structure-Aware Temporal Sequence Summarization via Statistical Window Merging
• A Hybrid Residual Dilated LSTM end Exponential Smoothing Model for Mid-Term Electric Load Forecasting
• CATS-Linear: Classification Auxiliary Linear Model for Time Series Forecasting
• Inner-Instance Normalization for Time Series Forecasting
• Adjusting for Autocorrelated Errors in Neural Networks for Time Series
• TimeGazer: Temporal Modeling of Predictive Gaze Stabilization for AR Interaction
• Test-Time Efficient Pretrained Model Portfolios for Time Series Forecasting
• Noise or Signal? Deconstructing Contradictions and An Adaptive Remedy for Reversible Normalization in Time Series Forecasting
• Forecasting-based Biomedical Time-series Data Synthesis for Open Data and Robust AI
• Forking-Sequences: Statistically and Computationally Efficient Multi-Horizon Forecasting with Reduced Volatility
• TimeSeriesScientist: A General-Purpose AI Agent for Time Series Analysis
• Understanding Transformers for Time Series: Rank Structure, Flow-of-ranks, and Compressibility
• Integrating AI and Ensemble Forecasting: Explainable Materials Planning with Scorecards and Trend Insights for a Large-Scale Manufacturer
• TimeEmb: A Lightweight Static-Dynamic Disentanglement Framework for Time Series Forecasting
• ReNF: Rethinking the Design of Neural Long-Term Time Series Forecasters
• User-Centric Communication Service Provision for Edge-Assisted Mobile Augmented Reality
• Kairos: Towards Adaptive and Generalizable Time Series Foundation Models
• Uncovering Zero-Shot Generalization Gaps in Time-Series Foundation Models Using Real-World Videos
• DSAT-HD: Dual-Stream Adaptive Transformer with Hybrid Decomposition for Multivariate Time Series Forecasting
• IndexNet: Timestamp and Variable-Aware Modeling for Time Series Forecasting
• ProtoTS: Learning Hierarchical Prototypes for Explainable Time Series Forecasting
• From Noise to Laws: Regularized Time-Series Forecasting via Denoised Dynamic Graphs
• Learning Temporal Saliency for Time Series Forecasting with Cross-Scale Attention
• Improving accuracy in short mortality rate series: Exploring Multi-step Forecasting Approaches in Hybrid Systems
• Mind the Missing: Variable-Aware Representation Learning for Irregular EHR Time Series using Large Language Models
• Information Bottleneck Learning for Faithful Time Series Forecasting Explanations
• A More Realistic Evaluation of Cross-Frequency Transfer Learning and Foundation Forecasting Models
• Multi-scale Temporal Prediction via Incremental Generation and Multi-agent Collaboration
• DPANet: Dual Pyramid Attention Network for Multivariate Time Series Forecasting
• TFMAdapter: Lightweight Instance-Level Adaptation of Foundation Models for Forecasting with Covariates
• NHITS: Neural Hierarchical Interpolation for Time Series Forecasting
• AquaCast: Urban Water Dynamics Forecasting with Precipitation-Informed Multi-Input Transformer
• Neural ARFIMA model for forecasting BRIC exchange rates with long memory
• ARIES: Relation Assessment and Model Recommendation for Deep Time Series Forecasting
• Real-E: A Foundation Benchmark for Advancing Robust and Generalizable Electricity Forecasting
• Privacy Risks in Time Series Forecasting: User- and Record-Level Membership Inference
• One-Embedding-Fits-All: Efficient Zero-Shot Time Series Forecasting by a Model Zoo
• Cross-device Zero-shot Label Transfer via Alignment of Time Series Foundation Model Embeddings
• An algorithmic framework for the optimization of deep neural networks architectures and hyperparameters
• PAX-TS: Model-agnostic multi-granular explanations for time series forecasting via localized perturbations
• RegimeNAS: Regime-Aware Differentiable Architecture Search With Theoretical Guarantees for Financial Trading
• TimeMKG: Knowledge-Infused Causal Reasoning for Multivariate Time Series Modeling
• N-BEATS-MOE: N-BEATS with a Mixture-of-Experts Layer for Heterogeneous Time Series Forecasting
• Comparative Analysis of Time Series Foundation Models for Demographic Forecasting: Enhancing Predictive Accuracy in US Population Dynamics
• Effective Sample Size and Generalization Bounds for Temporal Networks
• EnergyPatchTST: Multi-scale Time Series Transformers with Uncertainty Estimation for Energy Forecasting
• Salt-Rock Creep Deformation Forecasting Using Deep Neural Networks and Analytical Models for Subsurface Energy Storage Applications
• FeDaL: Federated Dataset Learning for Time Series Foundation Models
• DMSC: Dynamic Multi-Scale Coordination Framework for Time Series Forecasting
• Signals, Concepts, and Laws: Toward Universal, Explainable Time-Series Forecasting
• Towards Measuring and Modeling Geometric Structures in Time Series Forecasting via Image Modality
• Explainability-Driven Feature Engineering for Mid-Term Electricity Load Forecasting in ERCOT's SCENT Region
• Adaptive Fuzzy Time Series Forecasting via Partially Asymmetric Convolution and Sub-Sliding Window Fusion
• U-Cast: Learning Hierarchical Structures for High-Dimensional Time Series Forecasting
• HIST: A Graph-based Framework for Stock Trend Forecasting via Mining Concept-Oriented Shared Information
• Time Series Information Visualization -- A Review of Approaches and Tools
• Interpretable Sequence Learning for COVID-19 Forecasting
• Long-term series forecasting with Query Selector -- efficient model of sparse attention
• ParallelTime: Dynamically Weighting the Balance of Short- and Long-Term Temporal Dependencies
• Fremer: Lightweight and Effective Frequency Transformer for Workload Forecasting in Cloud Services
• Multivariate Long-term Time Series Forecasting with Fourier Neural Filter
• Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality
• Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directions
• KARMA: A Multilevel Decomposition Hybrid Mamba Framework for Multivariate Long-Term Time Series Forecasting
• TAT: Temporal-Aligned Transformer for Multi-Horizon Peak Demand Forecasting
• Wavelet-Enhanced Neural ODE and Graph Attention for Interpretable Energy Forecasting
• Fourier Basis Mapping: A Time-Frequency Learning Framework for Time Series Forecasting
• A Projection-Based ARIMA Framework for Nonlinear Dynamics in Macroeconomic and Financial Time Series: Closed-Form Estimation and Rolling-Window Inference
• Investigating the accuracy of cross-learning time series forecasting methods
• Mask6D: Masked Pose Priors For 6D Object Pose Estimation
• Forecasting CPI inflation under economic policy and geopolitical uncertainties
• OmniVec2 -- A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning
• Probabilistic Time Series Forecasting with Implicit Quantile Networks
• Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives
• MamNet: A Novel Hybrid Model for Time-Series Forecasting and Frequency Pattern Analysis in Network Traffic
• UniCA: Unified Covariate Adaptation for Time Series Foundation Model
• ITFormer: Bridging Time Series and Natural Language for Multi-Modal QA with Large-Scale Multitask Dataset
• FAF: A Feature-Adaptive Framework for Few-Shot Time Series Forecasting
• AutoAI-TS: AutoAI for Time Series Forecasting
• MIRA: Medical Time Series Foundation Model for Real-World Health Data
• SKOLR: Structured Koopman Operator Linear RNN for Time-Series Forecasting
• Multi-Scale Finetuning for Encoder-based Time Series Foundation Models
• Leveraging External Factors in Household-Level Electrical Consumption Forecasting using Hypernetworks
• CoIFNet: A Unified Framework for Multivariate Time Series Forecasting with Missing Values
• Probabilistic Time Series Forecasting with Structured Shape and Temporal Diversity
• Forecast-Then-Optimize Deep Learning Methods
• A Review of the Long Horizon Forecasting Problem in Time Series Analysis
• Comparative Analysis of Modern Machine Learning Models for Retail Sales Forecasting
• The effects of retraining on the stability of global models in retail demand forecasting
• LETS Forecast: Learning Embedology for Time Series Forecasting
• TimeRecipe: A Time-Series Forecasting Recipe via Benchmarking Module Level Effectiveness
• Few-shot Learning for Time-series Forecasting
• The cost of ensembling: is it always worth combining?
• Meta-Learning Dynamics Forecasting Using Task Inference
• Zero-Shot Time Series Forecasting with Covariates via In-Context Learning
• Dynamic Modes as Time Representation for Spatiotemporal Forecasting
• Revisiting LLMs as Zero-Shot Time-Series Forecasters: Small Noise Can Break Large Models
• Probabilistic Forecasting for Building Energy Systems using Time-Series Foundation Models
• Generalisation Bounds of Zero-Shot Economic Forecasting using Time Series Foundation Models
• Hierarchical Forecast Reconciliation on Networks: A Network Flow Optimization Formulation
• Entanglement for Pattern Learning in Temporal Data with Logarithmic Complexity: Benchmarking on IBM Quantum Hardware
• Improving Time Series Forecasting via Instance-aware Post-hoc Revision
• TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learning
• Mamba Integrated with Physics Principles Masters Long-term Chaotic System Forecasting
• Multi-Modal View Enhanced Large Vision Models for Long-Term Time Series Forecasting
• Forecasting Multivariate Urban Data via Decomposition and Spatio-Temporal Graph Analysis
• CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations
• Multi-Source Dynamic Graph Learning for Compound-Flood Forecasting in Managed Coastal Systems
• FlowMixer: A Constrained Neural Architecture for Interpretable Spatiotemporal Forecasting
• Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model
• Deep Learning-Based Forecasting of Boarding Patient Counts to Address ED Overcrowding
• Enhancing LLMs for Time Series Forecasting via Structure-Guided Cross-Modal Alignment
• PPTNet: A Hybrid Periodic Pattern-Transformer Architecture for Traffic Flow Prediction and Congestion Identification
• Context-Aware Probabilistic Modeling with LLM for Multimodal Time Series Forecasting
• IISE PG&E Energy Analytics Challenge 2025: Hourly-Binned Regression Models Beat Transformers in Load Forecasting
• Foundation Time-Series AI Model for Realized Volatility Forecasting
• A Multi-scale Representation Learning Framework for Long-Term Time Series Forecasting
• DELPHYNE: A Pre-Trained Model for General and Financial Time Series
• Cluster-and-Conquer: A Framework For Time-Series Forecasting
• Hierarchically Regularized Deep Forecasting
• M5 accuracy competition: Results, findings, and conclusions
• Horizon-resolved Forecastability of Time Series via Auto Mutual Information
• How Effective are Large Time Series Models in Hydrology? A Study on Water Level Forecasting in Everglades
• Instance-wise Graph-based Framework for Multivariate Time Series Forecasting
• AIRA₂: Overcoming Bottlenecks in AI Research Agents
• Pretrained Time-Series Foundation Models for Financial Return Forecasting
• E4GEN: Event-level Explainable Extreme-Enhanced Time-series Generation
• Report the Floor: A Training-Free Conformal Interval Is a Mandatory Baseline for Probabilistic Time-Series Forecasting
• Neural forecasting at scale
• On the retraining frequency of global models in retail demand forecasting
• Symmetry-Constrained Forecasting of Periodically Correlated Energy Processes
• Training-Free Probabilistic Time-Series Forecasting with Conformal Seasonal Pools
• Pretraining Large Brain Language Model for Active BCI: Silent Speech
• Stabilizing chaotic dynamical system reproduction in reservoir computing
• IDOBE: Infectious Disease Outbreak forecasting Benchmark Ecosystem
• SynEnergy: Anomaly Semantic-Guided Diffusion for Synthetic Energy Data Generation
• CANet: ChronoAdaptive Network for Enhanced Long-Term Time Series Forecasting under Non-Stationarity
• Forecasting: theory and practice
• An Artificial Intelligence-Based Framework for Predicting Emergency Department Overcrowding: Development and Evaluation Study
• Energy-Aware Pattern Disentanglement: A Generalizable Pattern Assisted Architecture for Multi-task Time Series Analysis
• Foundation Models for Time Series: A Survey
• SCENT: Robust Spatiotemporal Learning for Continuous Scientific Data via Scalable Conditioned Neural Fields
• ms-Mamba: Multi-scale Mamba for Time-Series Forecasting
• Enhancing Time Series Forecasting via Multi-Level Text Alignment with LLMs
• Simultaneously Reconciled Quantile Forecasting of Hierarchically Related Time Series

Related

• N-BEATS-MOE: N-BEATS with a Mixture-of-Experts Layer for Heterogeneous Time Series Forecasting
• Deep Learning for Time Series Forecasting: Tutorial and Literature Survey
• A Comprehensive Survey of Deep Learning for Time Series Forecasting: Architectural Diversity and Open Challenges
• Neural forecasting at scale
• WEITS: A Wavelet-enhanced residual framework for interpretable time series forecasting