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

TL;DR

This paper introduces LSTNet, a deep neural network framework combining CNN, RNN, and autoregressive models to effectively capture both long-term and short-term patterns in multivariate time series forecasting, outperforming existing methods.

Contribution

The paper presents a novel deep learning architecture, LSTNet, that effectively models both long- and short-term dependencies in multivariate time series data, addressing limitations of traditional models.

Findings

LSTNet outperforms state-of-the-art baselines on real-world datasets.

Combining CNN, RNN, and autoregressive models improves forecasting accuracy.

The framework effectively captures complex temporal patterns.

Abstract

Multivariate time series forecasting is an important machine learning problem across many domains, including predictions of solar plant energy output, electricity consumption, and traffic jam situation. Temporal data arise in these real-world applications often involves a mixture of long-term and short-term patterns, for which traditional approaches such as Autoregressive models and Gaussian Process may fail. In this paper, we proposed a novel deep learning framework, namely Long- and Short-term Time-series network (LSTNet), to address this open challenge. LSTNet uses the Convolution Neural Network (CNN) and the Recurrent Neural Network (RNN) to extract short-term local dependency patterns among variables and to discover long-term patterns for time series trends. Furthermore, we leverage traditional autoregressive model to tackle the scale insensitive problem of the neural network model. In our evaluation on real-world data with complex mixtures of repetitive patterns, LSTNet achieved significant performance improvements over that of several state-of-the-art baseline methods. All the data and experiment codes are available online.