用原始数据和变化率联合训练，提升时序预测精度。
Differential Machine Learning for Time Series Prediction
- 同时学习原始序列与微分序列，捕捉动态规律
- 在混沌系统与金融数据上均优于主流模型
- 适合缺乏完整机理知识的复杂时序预测
由于固有的非线性及对初值敏感，准确预测时间序列极具挑战。本文提出一种基于微分学习的新方法，通过在原始时间序列及其微分序列上联合训练神经网络来增强预测能力。具体而言，我们构建了微分长短期记忆（Diff-LSTM）网络，利用共享的LSTM单元同时处理两种数据流，有效捕捉内在模式与时间动态。在Mackey-Glass、Lorenz和Rössler混沌时间序列以及来自ACI Worldwide Inc.的真实金融数据集上进行评估，结果表明，Diff-LSTM在短时与长时预测中均优于循环神经网络、卷积神经网络、双向及编码器-解码器结构的LSTM网络。该框架为在缺乏完整系统动力学知识的情况下提升时间序列预测性能提供了有效方案。
原文摘要 · Abstract (English)
Accurate time series prediction is challenging due to the inherent nonlinearity and sensitivity to initial conditions. We propose a novel approach that enhances neural network predictions through differential learning, which involves training models on both the original time series and its differential series. Specifically, we develop a differential long short-term memory (Diff-LSTM) network that uses a shared LSTM cell to simultaneously process both data streams, effectively capturing intrinsic patterns and temporal dynamics. Evaluated on the Mackey-Glass, Lorenz, and Rössler chaotic time series, as well as a real-world financial dataset from ACI Worldwide Inc., our results demonstrate that the Diff- LSTM network outperforms prevalent models such as recurrent neural networks, convolutional neural networks, and bidirectional and encoder-decoder LSTM networks in both short-term and long-term predictions. This framework offers a promising solution for enhancing time series prediction, even when comprehensive knowledge of the underlying dynamics of the time series is not fully available.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。