arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.07282stat.MLcs.LG
收藏

时间序列的自适应共形预测

Adaptive Conformal Predictions for Time Series

Eletrcicité De France R&D(法国电力公司研发部);INRIA Sophia-Antipolis(法国国家信息与自动化研究所索菲亚-昂蒂波利斯分所);CMAP, Ecole Polytechnique, IP Paris(巴黎综合理工学院应用数学与概率中心(CMAP));FiME;IDESP;Eletrcicité De France R&D, Palaiseau, France(法国电力公司研发部(帕莱索));INRIA Sophia-Antipolis, Montpellier, France(法国国家信息与自动化研究所索菲亚-昂蒂波利斯分所(蒙彼利埃));CMAP, Ecole Polytechnique, IP Paris, Palaiseau, France(巴黎综合理工学院应用数学与概率中心(CMAP,帕莱索));FiME, Palaiseau, France(FiME(帕莱索));IDESP, Montpellier, France(IDESP(蒙彼利埃))

Margaux Zaffran, Aymeric Dieuleveut, Olivier Féron, Yannig Goude, Julie Josse

摘要:

预测模型的不确定性量化在决策问题中至关重要。共形预测是一个普遍的、理论上合理的答案。但是,它需要可交换的数据,不包括时间序列。虽然最近的工作解决了这个问题,我们认为,自适应共形推理(ACI,吉布斯和坎德{è}s,2021年),发展的分布移位时间序列,是一个很好的程序时间序列一般的依赖。我们从理论上分析了在可交换和自回归的情况下,学习率对其效率的影响。我们提出了一种无参数的方法,AgACI,自适应地建立在ACI的基础上在线专家聚集。我们领导广泛的公平模拟对竞争的方法,主张ACI的使用时间序列。我们进行了一个真实的案例研究:电价预测。所提出的聚合算法提供了有效的预测区间日前预测。再现实验的所有代码和数据都是可用的。

英文摘要:

Uncertainty quantification of predictive models is crucial in decision-making problems. Conformal prediction is a general and theoretically sound answer. However, it requires exchangeable data, excluding time series. While recent works tackled this issue, we argue that Adaptive Conformal Inference (ACI, Gibbs and Cand{è}s, 2021), developed for distribution-shift time series, is a good procedure for time series with general dependency. We theoretically analyse the impact of the learning rate on its efficiency in the exchangeable and auto-regressive case. We propose a parameter-free method, AgACI, that adaptively builds upon ACI based on online expert aggregation. We lead extensive fair simulations against competing methods that advocate for ACI's use in time series. We conduct a real case study: electricity price forecasting. The proposed aggregation algorithm provides efficient prediction intervals for day-ahead forecasting. All the code and data to reproduce the experiments is made available.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑