A TCN-Linear Hybrid Model for Chaotic Time Series Forecasting

Mengjiao Wang,Fengtai Qin

DOI: https://doi.org/10.3390/e26060467

IF: 2.738

2024-05-30

Entropy

Abstract:The applications of deep learning and artificial intelligence have permeated daily life, with time series prediction emerging as a focal area of research due to its significance in data analysis. The evolution of deep learning methods for time series prediction has progressed from the Convolutional Neural Network (CNN) and the Recurrent Neural Network (RNN) to the recently popularized Transformer network. However, each of these methods has encountered specific issues. Recent studies have questioned the effectiveness of the self-attention mechanism in Transformers for time series prediction, prompting a reevaluation of approaches to LTSF (Long Time Series Forecasting) problems. To circumvent the limitations present in current models, this paper introduces a novel hybrid network, Temporal Convolutional Network-Linear (TCN-Linear), which leverages the temporal prediction capabilities of the Temporal Convolutional Network (TCN) to enhance the capacity of LSTF-Linear. Time series from three classical chaotic systems (Lorenz, Mackey–Glass, and Rossler) and real-world stock data serve as experimental datasets. Numerical simulation results indicate that, compared to classical networks and novel hybrid models, our model achieves the lowest RMSE, MAE, and MSE with the fewest training parameters, and its R2 value is the closest to 1.

physics, multidisciplinary

What problem does this paper attempt to address?

This paper proposes a solution to the problem of chaotic time series prediction. With the widespread application of deep learning and artificial intelligence in daily life, time series prediction has become an important research field. Despite the developments of Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and the recently popular Transformer network in time series prediction, they each have their specific issues. The effectiveness of the self-attention mechanism in Transformers for Long-Term Sequence Forecasting (LTSF) has been questioned, prompting researchers to reconsider the approaches to dealing with such problems. The paper introduces a new hybrid network model called TCN-Linear, which combines the strengths of Temporal Convolutional Networks (TCNs) to enhance the capabilities of the LSTF-Linear model. TCN utilizes its time prediction ability to handle time series data, while LSTF-Linear directly predicts future values through linear layers. The TCN-Linear model is constructed with Residual Block modules and a D-Linear network, which includes Dilated Causal Conv, WeightNorm layers, ReLU layers, and Dropout layers. This structure aims to address the long-term dependencies and seasonal patterns in time series data, applicable to various time series prediction tasks, and reduces training parameters and time due to its efficient parallel computing capabilities. In the experimental section, the paper evaluates the model using three classic chaotic systems (Lorenz, Mackey-Glass, and Rossler), as well as real stock data. The results show that compared to traditional networks and novel hybrid models, the TCN-Linear model performs best in terms of RMSE, MAE, and MSE indicators, with the fewest training parameters. Its R2 value approaches 1, demonstrating high prediction accuracy.

A TCN-Linear Hybrid Model for Chaotic Time Series Forecasting

High-efficiency Chaotic Time Series Prediction Based on Time Convolution Neural Network

Effective LSTMs with Seasonal-Trend Decomposition and Adaptive Learning and Niching-Based Backtracking Search Algorithm for Time Series Forecasting

Chaotic Time Series Prediction Based on Long Short-Term Memory Neural Networks

TCLN: A Transformer-based Conv-LSTM Network for Multivariate Time Series Forecasting

A novel hybrid deep learning time series forecasting model based on long-short-term patterns

CTF-former: A Novel Simplified Multi-Task Learning Strategy for Simultaneous Multivariate Chaotic Time Series Prediction

Multivariate Temporal Convolutional Network: A Deep Neural Networks Approach for Multivariate Time Series Forecasting

A Hybrid Model Based on a Two-Layer Decomposition Approach and an Optimized Neural Network for Chaotic Time Series Prediction

Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks

TLNets: Transformation Learning Networks for long-range time-series prediction

Chaotic Time Series Prediction Based on Transiently Chaotic Neural Network

A Dilated Convolution Network-Based LSTM Model for Multi-Step Prediction of Chaotic Time Series

Chaotic Time Series Prediction Using DTIGNet Based on Improved Temporal-Inception and GRU

CTFNet: Long-Sequence Time-Series Forecasting Based on Convolution and Time–Frequency Analysis

A chaotic time series combined prediction model for improving trend lagging

Forecasting stock volatility and value-at-risk based on temporal convolutional networks

A Temporal Linear Network for Time Series Forecasting

Chaotic Time Series Prediction Using LSTM with CEEMDAN

Chaotic Time Series Forecasting Based on Dynamic Recurrent Neural Networks Model

Predicting Chaotic System Behavior using Machine Learning Techniques