DAD4TS: Data-Augmentation-Oriented Diffusion Model for Time-Series Forecasting with Small-Scale Data

2026-05-18

AI Summary

This study addresses the challenge of limited data augmentation efficacy in small-scale time series, which constrains predictive model performance. To overcome this, the authors propose DAD4TS, a novel data augmentation framework that integrates diffusion models with reinforcement learning to generate high-value samples guided within a geometric space, trained jointly with the forecasting model. Innovatively replacing the conventional variational autoencoder (VAE) with a mathematical projection better suited to small time series, DAD4TS embeds reinforcement learning into the diffusion process to optimize sample quality. Extensive experiments across six real-world datasets and eight forecasting models demonstrate that DAD4TS significantly outperforms seven baseline methods on five datasets, consistently enhancing prediction accuracy.

Abstract

Small-scale data is a critical problem in time-series forecasting tasks. Data augmentation is an effective strategy for this task, but it has a limitation in generating meaningful data. To address this limitation, we propose DAD4TS, a diffusion-model-based data augmentation method with reinforcement learning, designed for time-series forecasting with small-scale data. In DAD4TS, a data generator is simultaneously trained with a time-series model and controlled by a reinforcement learning model to efficiently generate samples that improve the forecast accuracy of the time-series model. To support small-scale data, we use mathematical methods instead of conventional VAE methods to train the diffusion model by projecting the time-series data into the geometric space. We validated the effectiveness of DAD4TS with seven comparative methods through qualitative and quantitative experiments on six real-world datasets and eight time-series models. As a result, DAD4TS was validated on five datasets.

Problem

• time-series forecasting
• small-scale data
• data augmentation
• meaningful data generation

Innovation

• diffusion model
• data augmentation
• reinforcement learning
• time-series forecasting
• small-scale data