M-3

The M3 dataset is a comprehensive benchmark for time series forecasting that contains a diverse collection of 3003 time series from various domains, used to evaluate the performance of forecasting methods.

Papers using M-3 (7)

One-for-All: A Lightweight Stabilized and Parameter-Efficient Pre-trained LLM for Time Series Forecasting2026Forking-Sequences2025N-BEATS: Neural basis expansion analysis for interpretable time series forecasting2019 · 517 citesPerformance of Deep Learning models with transfer learning for multiple-step-ahead forecasts in monthly time series2022 · 11 citesImproving forecasting by subsampling seasonal time series2021A novel weighted approach for time series forecasting based on visibility graph2021Fast Gibbs sampling for the local and global trend Bayesian exponential smoothing model2024