arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1810.02406stat.MLcs.LG
收藏

高维问题中的投影推理:预测与特征选择

Projective Inference in High-dimensional Problems: Prediction and Feature Selection

Helsinki Institute of Information Technology (HIIT)(赫尔辛基信息技术研究所(HIIT));Aalto University(阿尔托大学)

Juho Piironen, Markus Paasiniemi, Aki Vehtari

摘要:

讨论了高维数据下广义线性模型的预测推理和特征选择问题。我们认为,在许多情况下,人们可以从理论上合理的两阶段方法的决定中受益:首先,构建一个可能的非稀疏模型,预测良好,然后找到一个最小的子集的特征,表征预测。在第一步中建立的模型称为\emph{reference model},在后一步中的操作称为预测性\emph{projection}。这种方法的主要特点是,它找到了稀疏性和预测精度之间的一个很好的折衷,增益来自于利用所有可用的信息,包括先验知识和来自遗漏的特征。我们回顾了几种遵循这一原则的方法,并提供了新的方法学贡献。我们提出了一种新的投影技术,统一了现有的两种技术,是准确和快速计算。我们还提出了一种使用快速留一交叉验证来评估特征选择过程的方法,该方法允许简单直观的模型大小选择。此外,我们证明了一个定理,有助于理解的条件下,投影的方法可能是有益的。通过几个模拟和现实世界的例子说明的好处。

英文摘要:

This paper discusses predictive inference and feature selection for generalized linear models with scarce but high-dimensional data. We argue that in many cases one can benefit from a decision theoretically justified two-stage approach: first, construct a possibly non-sparse model that predicts well, and then find a minimal subset of features that characterize the predictions. The model built in the first step is referred to as the \emph{reference model} and the operation during the latter step as predictive \emph{projection}. The key characteristic of this approach is that it finds an excellent tradeoff between sparsity and predictive accuracy, and the gain comes from utilizing all available information including prior and that coming from the left out features. We review several methods that follow this principle and provide novel methodological contributions. We present a new projection technique that unifies two existing techniques and is both accurate and fast to compute. We also propose a way of evaluating the feature selection process using fast leave-one-out cross-validation that allows for easy and intuitive model size selection. Furthermore, we prove a theorem that helps to understand the conditions under which the projective approach could be beneficial. The benefits are illustrated via several simulated and real world examples.

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。

↑