TimeCatcher: A Variational Framework for Volatility-Aware Forecasting of Non-Stationary Time Series

2026-01-28

AI Summary

This work addresses the limitations of existing lightweight MLP-based models in long-term forecasting of highly non-stationary time series—such as network traffic—where reliance on local stationarity assumptions impedes the capture of abrupt fluctuations. To overcome this, we propose TimeCatcher, a novel framework that uniquely integrates volatility-aware mechanisms with variational modeling. Specifically, a variational encoder extracts latent dynamic patterns from historical data, while a dedicated volatility-aware enhancement module identifies and amplifies critical local changes, thereby circumventing the dependence on stationarity inherent in conventional linear architectures. Evaluated across nine real-world datasets spanning traffic, finance, energy, and weather domains, TimeCatcher consistently outperforms state-of-the-art methods, demonstrating superior robustness and accuracy—particularly in high-volatility scenarios requiring long-horizon predictions.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Large Vision Models
• Natural Language Processing: (Large) Language Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Recent lightweight MLP-based models have achieved strong performance in time series forecasting by capturing stable trends and seasonal patterns. However, their effectiveness hinges on an implicit assumption of local stationarity assumption, making them prone to errors in long-term forecasting of highly non-stationary series, especially when abrupt fluctuations occur, a common challenge in domains like web traffic monitoring. To overcome this limitation, we propose TimeCatcher, a novel Volatility-Aware Variational Forecasting framework. TimeCatcher extends linear architectures with a variational encoder to capture latent dynamic patterns hidden in historical data and a volatility-aware enhancement mechanism to detect and amplify significant local variations. Experiments on nine real-world datasets from traffic, financial, energy, and weather domains show that TimeCatcher consistently outperforms state-of-the-art baselines, with particularly large improvements in long-term forecasting scenarios characterized by high volatility and sudden fluctuations. Our code is available at https://github.com/ColaPrinceCHEN/TimeCatcher.

Problem

• non-stationary time series
• volatility
• long-term forecasting
• abrupt fluctuations
• time series forecasting

Innovation

• Volatility-Aware
• Variational Forecasting
• Non-Stationary Time Series
• Long-Term Forecasting
• Latent Dynamic Patterns