时间序列ARIMA模型应用于颅内压监测初步探讨

Preliminary Study of Applying Auto Regressive Integrated Moving Average Model (a Mathematical Model) to Monitoring Intracranial Hypertension

目的探讨时间序列自回归单整移动平均(ARIMA)模型在颅内压(ICP)监测中应用的可能性及其在该领域的应用前景。方法对11例急性颅内压增高患者硬脑膜外留置ICP监测探头,术后连续20min监测并记录ICP值,应用SAS/ETS软件对每个患者ICP数据分别建立ARIMA时间序列模型,拟合ICP变化过程。结果所建ARIMA模型能够拟合ICP随时间的变化规律,预测误差范围可在正负4mmHg以内。结论ARIMA时间序列模型可用于ICP监测,具有潜在临床应用前景。

Objective To explore the practicability of the application of auto regressive integrated moving average（ARIMA） model to monitoring intracranial pressure（ICP） and its application prospects in this field. Methods ICP was continuously recorded for 20 minutes by an epidural sensor in 11 patients with acute intracranial hypertension （ICH）. The ICP time series data of each patient with ICH were used to establish ARIMA model with SAS/ETS software. The process of variation the in ICP was imitated. Results The established ARIMA model can imitate the regulation of the variation in ICP along with time. The predictive errors range was within ±4mmHg. Conclusions The model of ARIMA is ideal one for monitoring ICP, and it is hopeful of clinical potential use in the future.