4.3 Article

Approximate leave-future-out cross-validation for Bayesian time series models

出版社

TAYLOR & FRANCIS LTD
DOI: 10.1080/00949655.2020.1783262

关键词

Time series analysis; cross-Validation; Bayesian inference; pareto Smoothed importance sampling

资金

  1. Academy of Finland [298742, 313122]
  2. Technology Industries of Finland Centennial Foundation [70007503]
  3. Academy of Finland (AKA) [313122, 313122, 298742] Funding Source: Academy of Finland (AKA)

向作者/读者索取更多资源

One of the common goals of time series analysis is to use the observed series to inform predictions for future observations. In the absence of any actual new data to predict, cross-validation can be used to estimate a model's future predictive accuracy, for instance, for the purpose of model comparison or selection. Exact cross-validation for Bayesian models is often computationally expensive, but approximate cross-validation methods have been developed, most notably methods for leave-one-out cross-validation (LOO-CV). If the actual prediction task is to predict the future given the past, LOO-CV provides an overly optimistic estimate because the information from future observations is available to influence predictions of the past. To properly account for the time series structure, we can use leave-future-out cross-validation (LFO-CV). Like exact LOO-CV, exact LFO-CV requires refitting the model many times to different subsets of the data. Using Pareto smoothed importance sampling, we propose a method for approximating exact LFO-CV that drastically reduces the computational costs while also providing informative diagnostics about the quality of the approximation.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.3
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据