研究人员引入了一种新颖的方法来评估用于单变量时间序列预测的机器学习模型中的滞后相关性。该方法利用了诸如Ghost variables和Shapley values等框架，并结合了加性重要性度量来定义自相关函数（auto-relevance）和部分自相关函数（partial auto-relevance）。此外，还提出了一种新技术，用同一模型生成的一步预测来替换基于联盟的方法中的缺失特征。通过模拟和真实世界数据分析，并利用季节性ARMA系列模型和循环神经网络（recurrent neural networks）的模型，证明了这些方法的有效性。 AI
影响 引入了时间序列预测中特征相关性的新技术，有望提高模型的可解释性和性能。
排序理由 该集群包含一篇详细介绍时间序列分析新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
We propose a model agnostic methodology to measure lag relevance in machine learning forecasting models applied to univariate time series. Particularly, we are working in the context of time series using the frameworks of Ghost variables and Shapley values, together with additive…
arXiv:2607.01959v1 Announce Type: new Abstract: We propose a model agnostic methodology to measure lag relevance in machine learning forecasting models applied to univariate time series. Particularly, we are working in the context of time series using the frameworks of Ghost vari…
We propose a model agnostic methodology to measure lag relevance in machine learning forecasting models applied to univariate time series. Particularly, we are working in the context of time series using the frameworks of Ghost variables and Shapley values, together with additive…