What question did this study set out to answer?

This research aims to enhance forecasting accuracy of univariate time series by addressing challenges like noise and non-stationarity.

March 14, 2026Scientific ReportsOpen Access

MultiScaleWave: a wavelet-based multiscale framework for univariate time series forecasting

View Full Paper

Ask AI

Bookmark

Authors

MultiScaleWave demonstrates improved forecasting performance in univariate time series, suggesting a robust approach for real-world data challenges.

This research aims to enhance forecasting accuracy of univariate time series by addressing challenges like noise and non-stationarity.
Proposed a deep learning framework, MultiScaleWave, for time series forecasting.
Applied multi-level discrete wavelet transforms to decompose time series into multiscale components.
Modeled each temporal component with a granularity-adaptive module and fused outputs for forecasting.
MultiScaleWave outperformed competitive baselines on benchmark datasets.
Achieved superior forecasting performance, validating its effectiveness and generalizability.

Zheng et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

MultiScaleWave demonstrates improved forecasting performance in univariate time series, suggesting a robust approach for real-world data challenges.