缺失协变量的统计、机器学习与深度学习方法:面向阿尔茨海默病研究的实证评估
Statistical, Machine Learning and Deep Learning Methods for Missing Covariates: An Empirical Evaluation with Applications to Alzheimer's Disease Research
Yiqing Pan · Sharon X. Xie · Shujie Ma · Angela L. Jefferson · Panpan Zhang
中文摘要
生物医学研究中缺失数据常见,尤其当生物标志物昂贵或需侵入性操作获取时,如阿尔茨海默病中的脑脊液指标。尽管从传统统计方法到现代机器学习(ML)与深度学习(DL)算法存在多种缺失数据处理方法,其在不同缺失机制下的相对性能仍缺乏充分理解。本研究对八种缺失协变量处理方法进行实证评估,包括完整案例分析、多重插补、基于ML的技术,以及多层感知机、变分自编码器和生成对抗网络等DL框架。总结其概念基础、实际实现考量、优势与局限。通过大量模拟,评估不同缺失机制与缺失比例下的估计偏差、效率、标准误准确性、置信区间覆盖率和计算负担。基于Vanderbilt Memory and Aging Project的应用进一步说明不同方法在真实生物医学场景回归分析中的影响。最后给出方法选择的实用建议,并指出有待未来研究的剩余挑战。
关键要点
- 01比较八种缺失协变量处理方法,涵盖完整案例分析、多重插补、ML与DL框架
- 02通过模拟评估偏差、效率、标准误、置信区间覆盖与计算负担
- 03基于Vanderbilt Memory and Aging Project展示真实生物医学回归分析影响
- 04总结各方法概念基础、实现考量、优势与局限并给出选择建议
- 05指出不同缺失机制下方法性能理解不足,仍有待研究挑战
解读
尚无解读。
原始英文摘要
arXiv:2611.25219v1 Announce Type: new Abstract: Missing data are common in biomedical research, particularly when biomarkers are costly or require invasive procedures to obtain, such as cerebrospinal fluid measures in Alzheimer's disease. Although a wide range of missing data methods exists, from traditional statistical approaches to modern machine learning (ML) and deep learning (DL) algorithms, their comparative performance under different missingness mechanisms remains insufficiently understood. In this study, we provide an empirical evaluation of eight methods for handling missing covariates, including complete-case analysis, multiple imputation, ML-based techniques, and DL frameworks such as multilayer perceptrons, variational autoencoders, and generative adversarial networks. We summarize their conceptual foundations, practical implementation considerations, strengths, and limitations. Through extensive simulations, we assess estimation bias, efficiency, standard error accuracy, confidence interval coverage, and computational burden across different missingness mechanisms and proportions. An application to the Vanderbilt Memory and Aging Project further illustrates the implications of different methods for regression analysis in a real-world biomedical setting. Finally, we provide practical recommendations for method selection and highlight remaining challenges that warrant future investigation.