Intracranial Pressure Forecasting in Children Using Dynamic Averaging of Time Series Data
Intracranial Pressure Forecasting in Children Using Dynamic Averaging of Time Series Data
复制标题
DOI：
10.3390/forecast1010004
复制
发表时间：
2019-12-01
期刊：
影响因子：
3
通讯作者：
Rasheed, Khaled
中科院分区：
文献类型：
--
作者：
Farhadi, Akram;Chern, Joshua J.;Rasheed, Khaled
颅内压升高（ICP）是一种严重的，往往危及生命的疾病。如果增加的压力推动关键的大脑结构和血管，它可能导致严重的永久性问题，甚至死亡。在这项研究中，我们提出了一种新的回归模型来预测ICP发作的儿童，提前30分钟，通过使用动态特性的连续颅内压，生命体征和药物治疗在过去的两个小时。分析了血压、呼吸频率、心率等生理参数与颅内压的相关性。使用线性回归、Lasso回归、支持向量机和随机森林算法来预测记录的ICP的接下来30分钟。最后，基于生命体征、药物和ICP创建动态特征。报告了血压与ICP（0.2）之间的弱相关性。在过去的两个小时内，通过使用给定的药物变量，随机森林模型的均方根误差（RMSE）从1.6%下降到0.89%。随机森林回归模型预测结果与实验值的相关系数为0.99。
Increased Intracranial Pressure (ICP) is a serious and often life-threatening condition. If the increased pressure pushes on critical brain structures and blood vessels, it can lead to serious permanent problems or even death. In this study, we propose a novel regression model to forecast ICP episodes in children, 30 min in advance, by using the dynamic characteristics of continuous intracranial pressure, vitals and medications during the last two hours. The correlation between physiological parameters, including blood pressure, respiratory rate, heart rate and the ICP, is analyzed. Linear regression, Lasso regression, support vector machine and random forest algorithms are used to forecast the next 30 min of the recorded ICP. Finally, dynamic features are created based on vitals, medications and the ICP. The weak correlation between blood pressure and the ICP (0.2) is reported. The Root-Mean-Square Error (RMSE) of the random forest model decreased from 1.6 to 0.89% by using the given medication variables in the last two hours. The random forest regression gave an accurate model for the ICP forecast with 0.99 correlation between the forecast and experimental values.