IdealTSF: Can Non-Ideal Data Contribute to Enhancing the Performance of Time Series Forecasting Models?

TL;DR

IdealTSF introduces a novel framework that leverages non-ideal negative samples alongside positive data to improve time series forecasting, especially in noisy or low-quality data scenarios.

Contribution

The paper proposes the IdealTSF framework, which effectively utilizes negative samples through pretraining, training, and optimization to enhance forecasting accuracy.

Findings

Negative samples significantly improve model performance.

IdealTSF outperforms existing methods on noisy datasets.

The approach is effective for low-quality and anomaly-prone data.

Abstract

Deep learning has shown strong performance in time series forecasting tasks. However, issues such as missing values and anomalies in sequential data hinder its further development in prediction tasks. Previous research has primarily focused on extracting feature information from sequence data or addressing these suboptimal data as positive samples for knowledge transfer. A more effective approach would be to leverage these non-ideal negative samples to enhance event prediction. In response, this study highlights the advantages of non-ideal negative samples and proposes the IdealTSF framework, which integrates both ideal positive and negative samples for time series forecasting. IdealTSF consists of three progressive steps: pretraining, training, and optimization. It first pretrains the model by extracting knowledge from negative sample data, then transforms the sequence data into ideal positive samples during training. Additionally, a negative optimization mechanism with adversarial disturbances is applied. Extensive experiments demonstrate that negative sample data unlocks significant potential within the basic attention architecture for time series forecasting. Therefore, IdealTSF is particularly well-suited for applications with noisy samples or low-quality data.