Division Measure, Model & Manage Bioresponses, Katholieke Universiteit Leuven, Leuven, Belgium.
J Med Syst. 2010 Jun;34(3):229-39. doi: 10.1007/s10916-008-9234-9.
This work studies the impact of using dynamic information as features in a machine learning algorithm for the prediction task of classifying critically ill patients in two classes according to the time they need to reach a stable state after coronary bypass surgery: less or more than 9 h. On the basis of five physiological variables (heart rate, systolic arterial blood pressure, systolic pulmonary pressure, blood temperature and oxygen saturation), different dynamic features were extracted, namely the means and standard deviations at different moments in time, coefficients of multivariate autoregressive models and cepstral coefficients. These sets of features served subsequently as inputs for a Gaussian process and the prediction results were compared with the case where only admission data was used for the classification. The dynamic features, especially the cepstral coefficients (aROC: 0.749, Brier score: 0.206), resulted in higher performances when compared to static admission data (aROC: 0.547, Brier score: 0.247). The differences in performance are shown to be significant. In all cases, the Gaussian process classifier outperformed to logistic regression.
本研究旨在探讨在机器学习算法中使用动态信息作为特征对冠状动脉旁路手术后达到稳定状态所需时间（<9 小时或>9 小时）的两类危重症患者进行分类预测任务的影响。基于五个生理变量（心率、收缩压、肺动脉收缩压、血液温度和氧饱和度），提取了不同的动态特征，即不同时间点的平均值和标准差、多元自回归模型的系数和倒谱系数。这些特征集随后作为高斯过程的输入，将预测结果与仅使用入院数据进行分类的情况进行比较。与静态入院数据相比（aROC：0.547，Brier 得分：0.247），动态特征，特别是倒谱系数（aROC：0.749，Brier 得分：0.206）表现出更高的性能。性能差异具有统计学意义。在所有情况下，高斯过程分类器的表现均优于逻辑回归。
