标题： 大型数据中贝叶斯模型比较的留一交叉验证 显示英文标题

标题： Leave-One-Out Cross-Validation for Bayesian Model Comparison in Large Data

摘要： 最近，基于子采样和后验近似的模型评估新方法已被提出，用于将留一交叉验证（LOO）扩展到大型数据集。 尽管这些方法在估计单个模型的预测性能方面效果良好，但在模型比较中效果较差。 我们提出了一种高效的方法，通过结合快速近似LOO代理与使用差值估计器的精确LOO子采样来估计预测性能的差异，并提供了关于扩展特性的证明。 这种方法比之前的方法效率高几个数量级，同时也更适合模型比较。 显示英文摘要

摘要： Recently, new methods for model assessment, based on subsampling and posterior approximations, have been proposed for scaling leave-one-out cross-validation (LOO) to large datasets. Although these methods work well for estimating predictive performance for individual models, they are less powerful in model comparison. We propose an efficient method for estimating differences in predictive performance by combining fast approximate LOO surrogates with exact LOO subsampling using the difference estimator and supply proofs with regards to scaling characteristics. The resulting approach can be orders of magnitude more efficient than previous approaches, as well as being better suited to model comparison.