Abstract:Time series forecasting has become an increasingly popular research area due to its critical applications in various real-world domains such as traffic management, weather prediction, and financial analysis. Despite significant advancements, existing models face notable challenges, including the necessity of manual hyperparameter tuning for different datasets, and difficulty in effectively distinguishing signal from redundant features in data characterized by strong seasonality. These issues hinder the generalization and practical application of time series forecasting models. To solve this issues, we propose an innovative time series forecasting model TimeSieve designed to address these challenges. Our approach employs wavelet transforms to preprocess time series data, effectively capturing multi-scale features without the need for additional parameters or manual hyperparameter tuning. Additionally, we introduce the information bottleneck theory that filters out redundant features from both detail and approximation coefficients, retaining only the most predictive information. This combination reduces significantly improves the model's accuracy. Extensive experiments demonstrate that our model outperforms existing state-of-the-art methods on 70% of the datasets, achieving higher predictive accuracy and better generalization across diverse datasets. Our results validate the effectiveness of our approach in addressing the key challenges in time series forecasting, paving the way for more reliable and efficient predictive models in practical applications. The code for our model is available at <a class="link-external link-https" href="https://github.com/xll0328/TimeSieve" rel="external noopener nofollow">this https URL</a>.

Trend Feature Extraction Method for Time Series Based on Double Evaluation Factors of Important Points

Trend feature extraction method based on important points in time series

A hybrid important points identification for time series: Financial case

A novel piecewise linear segmentation for time series

Real-time Algorithm for Trend Analysis of Dynamic Data Streams

Integrating Data-Driven Segmentation, Local Feature Extraction and Fisher Kernel Encoding to Improve Time Series Classification

Statistical Probability Based Algorithm for Extracting Feature Points in 2-Dimensional Image

Trade Filtering Method for Trend Following Strategy Based on LSTM-extracted Feature and Machine Learning

A Topological Based Feature Extraction Method for the Stock Market

Trend extraction based on variable time window length median filter

Variable Step Algorithm for Sub-Trend Sequence Searching

A Novel Trend Symbolic Aggregate Approximation for Time Series

Parallel Extraction of Long-term Trends and Short-term Fluctuation Framework for Multivariate Time Series Forecasting

Learning Non-Stationary Time-Series with Dynamic Pattern Extractions

A Time Series Double Threshold Method for Pollution Events Detection in Drinking Water Using Three-Dimensional Fluorescence Spectroscopy

TimeSieve: Extracting Temporal Dynamics through Information Bottlenecks

A Study of Polynomial Fit-Based Methods for Qualitative Trend Analysis

Main Trend Extraction Based on Irregular Sampling Estimation and Its Application in Storage Volume of Internet Data Center

Efficient Algorithm for a Novel Pattern of Time Series

A New Qualitative Trend Analysis Algorithm Based on Global Polynomial Fit

RobustTrend: A Huber Loss with a Combined First and Second Order Difference Regularization for Time Series Trend Filtering