机器学习评估炎症生物标志物对老年西班牙裔成人队列认知障碍的预测价值

中文摘要 AI 辅助

小型临床表格数据集需要可解释的机器学习，因为深度学习通常不切实际，而集成模型可能难以检查。一个关键的陷阱是统计显著性并不一定意味着预测效用。利用巴拿马老龄化研究倡议——健康差异（PARI-HD）队列（n=165）的数据，我们实现了一个无泄漏的阈值似然伯努利/分类朴素贝叶斯（BNB/CNB）分类器。在每个训练折内，每个连续预测变量被简化为一个监督的卡方派生状态，而收入通过分类似然进入模型。所有数据依赖步骤均在重复分层10折交叉验证（30次重复）内执行。人口统计学基线实现了ROC-AUC为0.630 ± 0.017。I-309（CCL1）是主要的增量特征，使AUC增加了0.110，配对的DeLong检验在100%的重复中产生p<0.05。在预先指定的主要分析中，I-309产生了固定分区的DeLong p=0.0018，并通过200个随机分区评估稳健性，其中中位p值为0.0011。在18个候选标记物的探索性家族中，I-309在冻结分区上实现了Benjamini-Hochberg调整的q=0.032，并在85%的随机分区中满足q<0.05，而其他标记物均未表现出可靠的增量预测价值。由于拟合模型是一个可检查的阈值和类条件概率表，这些结果将I-309/CCL1确定为用于表格预测认知障碍的可解释候选特征，有待外部验证。

英文摘要

Small clinical tabular datasets require interpretable machine learning because deep learning is often impractical and ensemble models can be difficult to inspect. A key pitfall is that statistical significance does not necessarily imply predictive utility. Using data from the Panama Aging Research Initiative--Health Disparities (PARI-HD) cohort (n=165), we implemented a leakage-safe threshold-likelihood Bernoulli/Categorical Naive Bayes (BNB/CNB) classifier. Within every training fold, each continuous predictor was reduced to a supervised chi-square-derived state, while income entered the model through a categorical likelihood. All data-dependent steps were performed within repeated stratified 10-fold cross-validation with 30 repeats. The demographic baseline achieved a ROC-AUC of 0.630 +/- 0.017. I-309 (CCL1) was the dominant incremental feature, increasing AUC by 0.110, with paired DeLong tests yielding p<0.05 in 100% of repeats. In the pre-specified primary analysis, I-309 produced a fixed-partition DeLong p=0.0018, with robustness assessed across 200 random partitions, where the median p-value was 0.0011. Within the exploratory family of 18 candidate markers, I-309 achieved a Benjamini-Hochberg-adjusted q=0.032 on the frozen partition and satisfied q<0.05 in 85% of random partitions, whereas no other marker demonstrated reliable incremental predictive value. Because the fitted model is an inspectable table of thresholds and class-conditional probabilities, these results identify I-309/CCL1 as an interpretable candidate feature for tabular prediction of cognitive impairment, pending external validation.

发表机构

• Worcester Polytechnic Institute（伍斯特理工学院）
• Centro de Vacunación e Investigación (CEVAXIN)（疫苗接种与研究中心（CEVAXIN））
• Universidad Tecnológica de Panamá（巴拿马理工大学）
• Instituto de Investigaciones Científicas y Servicios de Alta Tecnología (INDICASAT AIP)（高科技科学研究与服务研究所（INDICASAT AIP））
• Secretaría Nacional de Ciencia, Tecnología e Innovación (SENACYT)（国家科学、技术与创新秘书处（SENACYT））
• Universidad Católica Santa María La Antigua（圣玛丽亚拉安提瓜天主教大学）

机构由 AI 辅助整理，请以论文原文为准。