基于波形特征提取与支持向量机分类的颅内压增高预测研究
|
摘要
在神经内、外科，颅内压增高是一种常见的危重病症。颅内压增高能够导致脑灌注压降低与脑血流减少，造成脑组织缺血缺氧，甚至可能造成脑组织移位并产生脑疝。当临床医护人员发现患者发生颅内压增高之后才准备医护处理时，由于存在发现不及时、病情进展迅速、药物发挥作用时间较长等因素的影响，可能会错失最佳治疗时机，影响治疗效果，导致较高的颅脑伤亡率。
目前临床监护中尚缺乏有效的提前提示颅内压增高的报警设备。一个有效的颅内压增高预测方法能够提前提醒医护人员并使医护人员有足够的准备时间降低颅内压，防止病情恶化，避免脑疝和死亡等危重症状的出现。鉴于颅内压管理在预防继发损伤和提高患者预后的重要性，一个基于计算机自动分析的颅内压增高提前报警算法对于颅内压监护和医护处理具有重要意义。
随着信号处理技术的发展，许多研究人员提出了基于信号处理的自动颅内压增高预测方法。比较传统的方法主要建立在设置固定阈值和信号的独立性假设的基础上。自从90年代开始，部分学者考虑到信号内部的自关联性，对颅内压信号建立了时间序列模型，例如自回归模型。随着机器学习理论在诸多领域得到成功的应用与发展，近年来机器学习方法也被运用到颅内压趋势分析的研究中，例如人工神经网络等。还有学者将多种高级信号处理方法结合起来对颅内压信号趋势进行预测，例如小波分析，卡尔曼滤波和近似熵等方法。上述研究都将颅内压信号视为时间序列并将一段时间内的信号均值作为预测变量，忽略了每个心动周期内信号的动力学特性；而且颅内压波动是一个非线性和非平稳过程，科学界仍然缺乏关于颅脑调节机制的基本知识。这使得上述模型与方法不能准确的反应颅内压波动的机理，预测效果不够理想，未能被临床采用。然而，大量研究表明颅内压的单波波形特征与颅内压信号的趋势有关，同时这些波形特征也能够反应重要的颅内生理病理信息，如脑顺应性与脑血流自动调节能力。
鉴于颅内压单波波形特征与颅脑生理病理情况的高度相关性，本文提出一个新的对颅内压增高提前5分钟进行预测的方案。该方案的实施步骤如下：①设计一个逐拍分割算法将颅内压信号逐拍分割为单波信号；②设计一个波峰识别算法识别单波信号的特征子峰并提取波形特征指标；③构建一个基于支持向量机的二类分类系统，将颅内压单波波形特征指标作为分类系统的输入变量，系统的分类结果即对应着颅内压增高/颅内压未增高。具体地说，该分类系统首先利用一个具有全局搜索能力的优化特征选择算法，差分进化算法，自动选择最优的指标组合作为特征向量，利用Wrapper方案进行优化特征指标选择；然后将优化的特征向量作为输入变量，采用支持向量机作为分类器对波形特征进行分类。对优化特征向量的分类结果即对应着颅内压增高/颅内压未增高。
本论文的主要研究成果如下：
①提出一个新的颅内压信号逐拍分割算法。在当前的临床和科研工作中，主要利用与颅内压信号同步记录的其他信号的特征点分割颅内压的单波波形，这种方法在很多情形下不适用。本研究借鉴经典的图像匹配算法—形状上下文的思想实现了适合一维生理准周期信号的描述算子—波形上下文；本算法利用波形上下文提取给定点的波形特征，然后利用模板匹配法实现每个单波起搏点的检测。在没有同步记录的其他信号的情况下该算法可以实现连续颅内压信号的单波波形分割。
②提出一个新的颅内压单波波形特征提取算法。该算法利用波形上下文提取单波波形的形态特征，然后利用分类器—支持向量机对该特征进行分类，最终识别颅内压单波波形中的三个特征子峰以及峰峰值，潜伏期和收缩期斜率等波形特征指标。
③提出一个基于特征选择与支持向量机分类的颅内压增高预测系统。对于给定的信号，本系统在颅内压增高发生之前5分钟判断该信号是否为增高前段（颅内压增高）/平稳段（颅内压未增高）。该系统首先利用差分进化算法选择特征指标组合，利用Wrapper方案（差分进化算法进行特征选择，支持向量机作为分类器，敏感度与阳性预测值的均值作为决策函数）评估指标组合对分类的有效性并确定最优指标组合，然后将该指标组合输入支持向量机进行分类。分类器输出分类结果即对应着颅内压增高/颅内压未增高。
④本文提出的颅内压增高预测系统的一个重要特点是采用分类的方式，而非预测颅内压变化值的方式实现对颅内压增高的预测。对于给定的信号，本系统将之区分为增高前段和平稳段。当某段信号被区分为增高前段时即意味着5分钟以后将会产生颅内压增高。当某段信号被区分为平稳段时意味着5分钟以后将不会产生颅内压增高。通过这种分类的方式，将波形特征与颅内压增高直接联系起来。通过检测波形特征的变化判断是否将要发生颅内压增高，同时也为开展波形特征与颅脑病理情况的相关性开辟了一个新的研究途径。
最后，本文利用临床数据对上述方案进行了验证。首先手工标定颅内压信号的每个单波的起搏点和三个特征子峰，利用单波分割算法与单波波形特征提取算法分别依次识别单波起搏点和单波特征子峰，然后利用量化的评估准则对上述两个算法的性能进行了验证。最后利用分类系统对颅内压增高进行预测并设计对照试验评估分类系统的预测效果。对照试验中该系统取得的敏感度为84%，特异度为96%。截止到目前为止，本次研究是相关研究领域中第一次利用颅内压单波波形特征实现预测跨度为5分钟的颅内压增高预测。初步的实验结果验证了该系统的有效性和潜在的实用性，为进一步的临床应用提供了实验基础。
In neurological settings, intracranial hypertension (ICH) is a commonly happenedcritical condition. ICH can cause the decrease of cerebral perfusion pressure (CPP) andreduce the cerebral blood flow (CBF), which leads to brain hypoxia-ischemia and evenbrain herniation. Normally when the ICH is observed, the nursing procedure can bestarted. Due to that the ICH may not be found timely, it usually progresses rapidly anddrugs take effect with a long time, the prompt and effective treatment that can stopprogress or shorten the duration and prevent complications from an already existingdisease process may be delayed.
In current clinical settings, there is short of effective alarm device for predictingthe occurring of ICH. This kind of device can warn the nursing staff the happening ofICH and let them be well prepared for preventing the increase of intracranial pressure(ICP) and secondary brain injury, improving the clinical outcome of the treatment,preventing the happening of brain herniation and lowering the mortality. Thus, anautomatic ICH alert algorithm is of great value for the ICP management and its nursingcare.
With the development of the signal processing technology, many ICH forecastingmethods have been proposed to alarm the ICH in advance. The conventional methodsinclude the setting up of a threshold and the independence hypothesis of the ICP signal.From90s, the autocorrelation of the ICP signal is taken into consideration and timeseries regression model is employed to deal with this problem. With the success of theapplication and development of machine learning technology in many fields, severalresearchers adopt the machine learning technology in the prediction of ICH in recentyears, e.g., the artificial neural network. And researchers combine several advancedsignal processing methods to attack this problem, such as the wavelet analysis, kalmanfiltering and approximate entropy. In these methods, the ICP signal is averaged and themean value is considered as the predictor. It ignores the dynamic characteristic of thesignal during the cardiac cycle. In addition, the ICP signal is a nonlinear andnonstationary process and the fundamental knowledge of the mechanism ofcerebrovascular regulation is insufficiency, which makes the above models and methodsnot good enough and not adopted in clinical practice.
However, recent research findings have shown that the configuration of thecharacteristic peaks and other waveform features of the pulse waveform have strongpositive correlation with the neurological conditions as well as the evolvement of theICP. They also can disclose important information of intracranial environment, such asintracranial compliance (IC) and cerebral autoregulation (CA).
In view of the high relation between the waveform feature and the brainpathophysiology, this paper proposes a novel scheme to judge if the ICH will occurafter5minutes. The procedure of forecasting the ICH of the proposed method is asfollows.①The continuous ICP signal is segmented to individual pulse by a new pulseonset detection algorithm.②The three characteristic peaks are identified by a peakrecognition algorithm and the morphological feature metrics are extracted.③Aclassification system is constructed based on the support vector machines (SVM). Thefeature metrics is the input of the classification system and the classification resultscorrespond to ICH/non-ICH. Specifically, the system first employs a global searchalgorithm, deferential evolution (DE) algorithm, to select the optimal featurecombinations and Wrapper scheme to perform the optimal feature selection. Then theselected optimal feature combinations are classified by the classifier SVM. The outputof the classifier corresponds to the ICH or non-ICH group.
The main research findings of this paper are as follows.
First, a novel algorithm is proposed to detect the pulse onset of continuous ICPsignal. There are on other methods existing to detect the pulse onset of continuous ICPsignal. Following the idea of a well known algorithm in image processing field, shapecontext, a descriptor, waveform descriptor, to describe the points on one-dimensionalphysiological signal is constructed. Then the waveform descriptor is employed toextract the feature of given points and the features are compared with a customizedtemplate and finally the onset whose feature is most similar with the template isidentified. After the onset is identified, the continuous ICP signal is segmented intoindividual pulse.
Second, a novel algorithm to recognize the characteristic peaks of ICP pulses isproposed. There are on other methods existing to recognize the characteristic peaks ofeach individual ICP pulse. The algorithm first employs the waveform descriptor toextract the features of the points on the ICP pulse, and then utilizes a classifier, SVM, toclassify the features and the three characteristic peaks of ICP pulses are identified. Finally, the morphological features such as latency, slope and peak-to-peak amplitudeare extracted.
Three, an ICH forecasting algorithm is proposed based on optimal feature selectionand SVM classification. For a given signal, the algorithm aims to discriminate betweenthe pre-ICH segment and stable segment5min before the occurring of the ICH. Thealgorithm is able to judge if the signal is ICH before5minutes to happen. It firstemploys the DE algorithm to select the optimal feature sub-set, Wrapper method (DEalgorithm for feature selection, SVM for classification and the mean value of thesensitivity and positive predictive value as the objective function) to evaluate theperformance of the selected sub-set and determine the optimal feature set and then usesSVM to classify the optimal feature to different classes. The output of the classifiercorresponds to the pre-ICH segment (ICH) or stable segment (non-ICH).
Four, one important feature of the proposed ICH forecasting system is theclassification way is adopted, rather than the prediction of the future value of the ICPsignal. For a given signal, the algorithm can discriminate between the pre-ICH segmentand stable segment. Once it is regarded as stable segment, it means that the ICH will nothappen in5minutes. Otherwise, once it is regarded as pre-ICH segment, it means thatthe ICH will happen in5minutes. The classification system sets up some directconnection between the morphological features and ICH, forecasts the ICH by detectingthe change of the morphological features. It provides a new window to explore therelation between the morphological features and neurological conditions.
Finally, the proposed method is validated. First, the onset and three characteristicpeaks are annotated manually and the onset detection and peak recognition algorithmare performed, respectively. The performance of these two algorithms is validated usingquantitative evaluation criteria. The ICH is forecasted by the SVM classification systemand the system performance is evaluated by control experiment. The sensitivityachieved by the system is84%and the specificity is96%. To our knowledge, this studyis the first attempt of ICH forecastion with a time span of5minutes. The preliminaryresults indicate its effectiveness and the potential practicality and provide experimentalfoundation for the clinical application.
目前临床监护中尚缺乏有效的提前提示颅内压增高的报警设备。一个有效的颅内压增高预测方法能够提前提醒医护人员并使医护人员有足够的准备时间降低颅内压，防止病情恶化，避免脑疝和死亡等危重症状的出现。鉴于颅内压管理在预防继发损伤和提高患者预后的重要性，一个基于计算机自动分析的颅内压增高提前报警算法对于颅内压监护和医护处理具有重要意义。
随着信号处理技术的发展，许多研究人员提出了基于信号处理的自动颅内压增高预测方法。比较传统的方法主要建立在设置固定阈值和信号的独立性假设的基础上。自从90年代开始，部分学者考虑到信号内部的自关联性，对颅内压信号建立了时间序列模型，例如自回归模型。随着机器学习理论在诸多领域得到成功的应用与发展，近年来机器学习方法也被运用到颅内压趋势分析的研究中，例如人工神经网络等。还有学者将多种高级信号处理方法结合起来对颅内压信号趋势进行预测，例如小波分析，卡尔曼滤波和近似熵等方法。上述研究都将颅内压信号视为时间序列并将一段时间内的信号均值作为预测变量，忽略了每个心动周期内信号的动力学特性；而且颅内压波动是一个非线性和非平稳过程，科学界仍然缺乏关于颅脑调节机制的基本知识。这使得上述模型与方法不能准确的反应颅内压波动的机理，预测效果不够理想，未能被临床采用。然而，大量研究表明颅内压的单波波形特征与颅内压信号的趋势有关，同时这些波形特征也能够反应重要的颅内生理病理信息，如脑顺应性与脑血流自动调节能力。
鉴于颅内压单波波形特征与颅脑生理病理情况的高度相关性，本文提出一个新的对颅内压增高提前5分钟进行预测的方案。该方案的实施步骤如下：①设计一个逐拍分割算法将颅内压信号逐拍分割为单波信号；②设计一个波峰识别算法识别单波信号的特征子峰并提取波形特征指标；③构建一个基于支持向量机的二类分类系统，将颅内压单波波形特征指标作为分类系统的输入变量，系统的分类结果即对应着颅内压增高/颅内压未增高。具体地说，该分类系统首先利用一个具有全局搜索能力的优化特征选择算法，差分进化算法，自动选择最优的指标组合作为特征向量，利用Wrapper方案进行优化特征指标选择；然后将优化的特征向量作为输入变量，采用支持向量机作为分类器对波形特征进行分类。对优化特征向量的分类结果即对应着颅内压增高/颅内压未增高。
本论文的主要研究成果如下：
①提出一个新的颅内压信号逐拍分割算法。在当前的临床和科研工作中，主要利用与颅内压信号同步记录的其他信号的特征点分割颅内压的单波波形，这种方法在很多情形下不适用。本研究借鉴经典的图像匹配算法—形状上下文的思想实现了适合一维生理准周期信号的描述算子—波形上下文；本算法利用波形上下文提取给定点的波形特征，然后利用模板匹配法实现每个单波起搏点的检测。在没有同步记录的其他信号的情况下该算法可以实现连续颅内压信号的单波波形分割。
②提出一个新的颅内压单波波形特征提取算法。该算法利用波形上下文提取单波波形的形态特征，然后利用分类器—支持向量机对该特征进行分类，最终识别颅内压单波波形中的三个特征子峰以及峰峰值，潜伏期和收缩期斜率等波形特征指标。
③提出一个基于特征选择与支持向量机分类的颅内压增高预测系统。对于给定的信号，本系统在颅内压增高发生之前5分钟判断该信号是否为增高前段（颅内压增高）/平稳段（颅内压未增高）。该系统首先利用差分进化算法选择特征指标组合，利用Wrapper方案（差分进化算法进行特征选择，支持向量机作为分类器，敏感度与阳性预测值的均值作为决策函数）评估指标组合对分类的有效性并确定最优指标组合，然后将该指标组合输入支持向量机进行分类。分类器输出分类结果即对应着颅内压增高/颅内压未增高。
④本文提出的颅内压增高预测系统的一个重要特点是采用分类的方式，而非预测颅内压变化值的方式实现对颅内压增高的预测。对于给定的信号，本系统将之区分为增高前段和平稳段。当某段信号被区分为增高前段时即意味着5分钟以后将会产生颅内压增高。当某段信号被区分为平稳段时意味着5分钟以后将不会产生颅内压增高。通过这种分类的方式，将波形特征与颅内压增高直接联系起来。通过检测波形特征的变化判断是否将要发生颅内压增高，同时也为开展波形特征与颅脑病理情况的相关性开辟了一个新的研究途径。
最后，本文利用临床数据对上述方案进行了验证。首先手工标定颅内压信号的每个单波的起搏点和三个特征子峰，利用单波分割算法与单波波形特征提取算法分别依次识别单波起搏点和单波特征子峰，然后利用量化的评估准则对上述两个算法的性能进行了验证。最后利用分类系统对颅内压增高进行预测并设计对照试验评估分类系统的预测效果。对照试验中该系统取得的敏感度为84%，特异度为96%。截止到目前为止，本次研究是相关研究领域中第一次利用颅内压单波波形特征实现预测跨度为5分钟的颅内压增高预测。初步的实验结果验证了该系统的有效性和潜在的实用性，为进一步的临床应用提供了实验基础。
In neurological settings, intracranial hypertension (ICH) is a commonly happenedcritical condition. ICH can cause the decrease of cerebral perfusion pressure (CPP) andreduce the cerebral blood flow (CBF), which leads to brain hypoxia-ischemia and evenbrain herniation. Normally when the ICH is observed, the nursing procedure can bestarted. Due to that the ICH may not be found timely, it usually progresses rapidly anddrugs take effect with a long time, the prompt and effective treatment that can stopprogress or shorten the duration and prevent complications from an already existingdisease process may be delayed.
In current clinical settings, there is short of effective alarm device for predictingthe occurring of ICH. This kind of device can warn the nursing staff the happening ofICH and let them be well prepared for preventing the increase of intracranial pressure(ICP) and secondary brain injury, improving the clinical outcome of the treatment,preventing the happening of brain herniation and lowering the mortality. Thus, anautomatic ICH alert algorithm is of great value for the ICP management and its nursingcare.
With the development of the signal processing technology, many ICH forecastingmethods have been proposed to alarm the ICH in advance. The conventional methodsinclude the setting up of a threshold and the independence hypothesis of the ICP signal.From90s, the autocorrelation of the ICP signal is taken into consideration and timeseries regression model is employed to deal with this problem. With the success of theapplication and development of machine learning technology in many fields, severalresearchers adopt the machine learning technology in the prediction of ICH in recentyears, e.g., the artificial neural network. And researchers combine several advancedsignal processing methods to attack this problem, such as the wavelet analysis, kalmanfiltering and approximate entropy. In these methods, the ICP signal is averaged and themean value is considered as the predictor. It ignores the dynamic characteristic of thesignal during the cardiac cycle. In addition, the ICP signal is a nonlinear andnonstationary process and the fundamental knowledge of the mechanism ofcerebrovascular regulation is insufficiency, which makes the above models and methodsnot good enough and not adopted in clinical practice.
However, recent research findings have shown that the configuration of thecharacteristic peaks and other waveform features of the pulse waveform have strongpositive correlation with the neurological conditions as well as the evolvement of theICP. They also can disclose important information of intracranial environment, such asintracranial compliance (IC) and cerebral autoregulation (CA).
In view of the high relation between the waveform feature and the brainpathophysiology, this paper proposes a novel scheme to judge if the ICH will occurafter5minutes. The procedure of forecasting the ICH of the proposed method is asfollows.①The continuous ICP signal is segmented to individual pulse by a new pulseonset detection algorithm.②The three characteristic peaks are identified by a peakrecognition algorithm and the morphological feature metrics are extracted.③Aclassification system is constructed based on the support vector machines (SVM). Thefeature metrics is the input of the classification system and the classification resultscorrespond to ICH/non-ICH. Specifically, the system first employs a global searchalgorithm, deferential evolution (DE) algorithm, to select the optimal featurecombinations and Wrapper scheme to perform the optimal feature selection. Then theselected optimal feature combinations are classified by the classifier SVM. The outputof the classifier corresponds to the ICH or non-ICH group.
The main research findings of this paper are as follows.
First, a novel algorithm is proposed to detect the pulse onset of continuous ICPsignal. There are on other methods existing to detect the pulse onset of continuous ICPsignal. Following the idea of a well known algorithm in image processing field, shapecontext, a descriptor, waveform descriptor, to describe the points on one-dimensionalphysiological signal is constructed. Then the waveform descriptor is employed toextract the feature of given points and the features are compared with a customizedtemplate and finally the onset whose feature is most similar with the template isidentified. After the onset is identified, the continuous ICP signal is segmented intoindividual pulse.
Second, a novel algorithm to recognize the characteristic peaks of ICP pulses isproposed. There are on other methods existing to recognize the characteristic peaks ofeach individual ICP pulse. The algorithm first employs the waveform descriptor toextract the features of the points on the ICP pulse, and then utilizes a classifier, SVM, toclassify the features and the three characteristic peaks of ICP pulses are identified. Finally, the morphological features such as latency, slope and peak-to-peak amplitudeare extracted.
Three, an ICH forecasting algorithm is proposed based on optimal feature selectionand SVM classification. For a given signal, the algorithm aims to discriminate betweenthe pre-ICH segment and stable segment5min before the occurring of the ICH. Thealgorithm is able to judge if the signal is ICH before5minutes to happen. It firstemploys the DE algorithm to select the optimal feature sub-set, Wrapper method (DEalgorithm for feature selection, SVM for classification and the mean value of thesensitivity and positive predictive value as the objective function) to evaluate theperformance of the selected sub-set and determine the optimal feature set and then usesSVM to classify the optimal feature to different classes. The output of the classifiercorresponds to the pre-ICH segment (ICH) or stable segment (non-ICH).
Four, one important feature of the proposed ICH forecasting system is theclassification way is adopted, rather than the prediction of the future value of the ICPsignal. For a given signal, the algorithm can discriminate between the pre-ICH segmentand stable segment. Once it is regarded as stable segment, it means that the ICH will nothappen in5minutes. Otherwise, once it is regarded as pre-ICH segment, it means thatthe ICH will happen in5minutes. The classification system sets up some directconnection between the morphological features and ICH, forecasts the ICH by detectingthe change of the morphological features. It provides a new window to explore therelation between the morphological features and neurological conditions.
Finally, the proposed method is validated. First, the onset and three characteristicpeaks are annotated manually and the onset detection and peak recognition algorithmare performed, respectively. The performance of these two algorithms is validated usingquantitative evaluation criteria. The ICH is forecasted by the SVM classification systemand the system performance is evaluated by control experiment. The sensitivityachieved by the system is84%and the specificity is96%. To our knowledge, this studyis the first attempt of ICH forecastion with a time span of5minutes. The preliminaryresults indicate its effectiveness and the potential practicality and provide experimentalfoundation for the clinical application.
引文
[1]陈维福,程远,马颖.高血压性脑出血外科治疗术式浅析[J].中国临床神经外科杂志,2003,8:(3)217-218.
[2]岳献芳,王立,周峰.颅内压变化引起的颅骨表面应变分析[J].北京科技大学学报,2006,28:(012)1143-1151.
[3] Jennum P, B rgesen SE. Intracranial pressure and obstructive sleep apnea[J]. Chest,1989,95:(2)279-283.
[4] Marmarou A, Anderson RL, Ward JD, Choi SC, Young HF, Eisenberg HM, Foulkes MA,Marshall LF, Jane JA. Impact of ICP instability and hypotension on outcome in patients withsevere head trauma[J]. Special Supplements,1991,75:(1S)59-66.
[5]钱素云,樊寻梅.急性颅内压增高与应激反应[J]. PEDIATRIC EMERGENCY MEDICINE,2003,10:(2).
[6]王群思.颅内压监测在危重症监护中的应用[J]. CHINESE PEDIATRIC EMERGENCYMEDICINE,2007,14:(3).
[7]李振举.颅脑损伤后脑血流变化及自由基反应对预后的影响[D],长春：吉林大学硕士学位论文.2007.
[8] Ohta H, Hadeishi H, Nemoto M, Kawamura S, Hinuma Y, Suzuki E. Transient Effect ofHyperbaric Oxygen on Cerebral Blood Flow and Intracranial Pressure[J].1990.
[9]杨勇,杨义,王任直.颅内压监测技术的进展[J].中国微侵袭神经外科杂志,2008,13:(6)284.
[10]苏治国,张赛.高渗盐水对脑水肿和颅内高压治疗作用的研究进展[J].武警医学院学报,2006,15:(006)607-610.
[11]徐姜定.颅内压增高监测及治疗进展[J].浙江临床医学,2003,5:(7)481-482.
[12] Guttag JJ, Ganeshapillai G. Methods to improve the signal quality of corruptedmulti-parameter physiological signals[D], Massachusetts Institute of Technology.2011.
[13] Hornero R, Aboy M, Abásolo D, McNames J, Goldstein B. Interpretation of approximateentropy: analysis of intracranial pressure approximate entropy during acute intracranialhypertension[J]. IEEE Transactions on Biomedical Engineering,2005,52:(10)1671-1680.
[14] Hamilton R, Xu P, Asgari S, Kasprowicz M, Vespa P, Bergsneider M, Hu X. Forecastingintracranial pressure elevation using pulse waveform morphology[C]. IEEE,2009,4331-4334.
[15] Wiklund L, Hok B, Stahl K, Jordeby-Jonsson A. Postanesthesia monitoring revisited:frequency of true and false alarms from different monitoring devices[J]. Journal of clinicalanesthesia,1994,6:(3)182-188.
[16] Lawless ST. Crying wolf: false alarms in a pediatric intensive care unit[J]. Critical caremedicine,1994,22:(6)981.
[17] Imhoff M, Bauer M, Gather U, Fried R. Pattern detection in intensive care monitoring timeseries with autoregressive models: Influence of the model order[J]. Biometrical Journal,2002,44:(6)746-761.
[18] Ursino M, Iezzi M, Stocchetti N. Intracranial pressure dynamics in patients with acute braindamage: a critical analysis with the aid of a mathematical model[J]. IEEE Transactions onBiomedical Engineering,1995,42:(6)529-540.
[19] Swiercz M, Mariak Z, Lewko J, Chojnacki K, Kozlowski A, Piekarski P. Neural networktechnique for detecting emergency states in neurosurgical patients[J]. Medical and BiologicalEngineering and Computing,1998,36:(6)717-722.
[20] Charbonnier S, Becq G, Biot L. On-line segmentation algorithm for continuously monitoreddata in intensive care units[J]. IEEE Transactions on Biomedical Engineering,2004,51:(3)484-492
[21] Charbonnier S, Gentil S. A trend-based alarm system to improve patient monitoring inintensive care units[J]. Control engineering practice,2007,15:(9)1039-1050.
[22] Hinkley DV. Inference about the change-point from cumulative sum tests[J]. Biometrika,1971,58:(3)509.
[23] Keogh E, Chu S, Hart D, Pazzani M. An online algorithm for segmenting time series[C]. inProceedings IEEE International Conference on Data Mining. IEEE,2001,289-296.
[24] Kirkness CJ, Mitchell PH, Burr RL, March KS, Newell DW. Intracranial pressure waveformanalysis: clinical and research implications[J]. The Journal of neuroscience nursing: journal ofthe American Association of Neuroscience Nurses,2000,32:(5)271.
[25] Swiercz M, Mariak Z, Krejza J, Lewko J, Szydlik P. Intracranial pressure processing withartificial neural networks: Prediction of ICP trends[J]. Acta neurochirurgica,2000,142:(4)401-406.
[26] Fan JY, Kirkness C, Vicini P, Burr R, Mitchell P. Intracranial Pressure WaveformMorphology and Intracranial Adaptive Capacity[J]. American Journal of Critical Care,2008,17:(6)545.
[27] Westhout FD, Paré LS, Delfino RJ, Cramer SC. Slope of the intracranial pressure waveformafter traumatic brain injury[J]. Surgical neurology,2008,70:(1)70-74.
[28] Hickey JV, Olson DWM, Turner DA. Intracranial pressure waveform analysis during rest andsuctioning[J]. Biological Research For Nursing,2009,11:(2)174.
[29] Contant CF. Intracranial pressure waveform indices in transient and refractory intracranialhypertension[J]. Journal of Neuroscience methods,1995,57:(1)15-25.
[30] Lundberg N. Continuous recording and control of ventricular fluid pressure in neurosurgicalpractice[J]. Acta psychiatrica Scandinavica. Supplementum,1960,36:(149)1.
[31] Czosnyka M, Smielewski P, Piechnik S, Schmidt EA, Al-Rawi PG, Kirkpatrick PJ, Pickard JD.Hemodynamic characterization of intracranial pressure plateau waves in head-injuredpatients[J]. Journal of neurosurgery,1999,91:(1)11-19.
[32] Szewczykowski J, Dytko P, Kunicki A, Korsak-Sliwka J, Sliwka S, Dziduszko J, AugustyniakB. Determination of critical ICP levels in neurosurgical patients: a statistical approach[J].Intracranial Pressure II,1975,392-393.
[33] Turner JM, McDowall DG, Gibson RM, Khaili H. Computer analysis of intracranial pressuremeasurements: Clinical value and nursing response[J]. Intracranial pressure III,1976,283-287.
[34] Price DJ, Dugdale RE, Mason J. The control of ICP using three asynchronous closed loops[J].Intracranial Pressure IV,1980,395-399.
[35] Portnoy HD, Chopp M. Cerebrospinal fluid pulse wave form analysis during hypercapnia andhypoxia[J]. Neurosurgery,1981,9:(1)14.
[36] McNames J, Crespo C, Bassale J, Aboy M, Elleby M, Lai S, Goldstein B. Sensitive precursorsto acute episodes of intracranial hypertension[C]. in The4th International Workshop inBiosignal Interpretation. Como, Italy,2002,303-306.
[37] Pincus SM. Approximate entropy as a measure of system complexity[J]. Proceedings of theNational Academy of Sciences,1991,88:(6)2297.
[38] Hornero R, Aboy M, Abasolo D, McNames J, Wakeland W, Goldstein B. Complex analysis ofintracranial hypertension using approximate entropy[J]. Critical care medicine,2006,34:(1)87.
[39] Hu X, Miller C, Vespa P, Bergsneider M. Adaptive computation of approximate entropy andits application in integrative analysis of irregularity of heart rate variability and intracranialpressure signals[J]. Medical engineering&physics,2008,30:(5)631-639.
[40] Hu X, Xu P, Asgari S, Vespa P, Bergsneider M. Forecasting ICP elevation based on prescientchanges of intracranial pressure waveform morphology[J]. IEEE Transactions on BiomedicalEngineering,2010,57:(5)1070-1078.
[41] Eide PK, Czosnyka M, Sorteberg W, Pickard JD, Smielewski P. Association betweenintracranial, arterial pulse pressure amplitudes and cerebral autoregulation in head injurypatients[J]. Neurological Research,2007,29:(6)578-582.
[42]姚旭,王晓丹,张玉玺,权文.特征选择方法综述[J].控制与决策,2012.
[43] Weston J, Mukherjee S, Chapelle O, Pontil M, Poggio T, Vapnik V. Feature selection forSVMs[J]. Advances in neural information processing systems,2001,668-674.
[44] Maxeiner H, Behnke M. Intracranial volume, brain volume, reserve volume andmorphological signs of increased intracranial pressure-A post-mortem analysis[J]. LegalMedicine,2008,10:(6)293-300.
[45]刘宏雨.侧脑室引流在68例脑室出血患者中的应用及护理[J].中华护理杂志,2005,40:(007)516-518.
[46] Shahsavari S, McKelvey T. Harmonics tracking of intracranial and arterial blood pressurewaves[C]. in30th Annual IEEE International Conference on Engineering in Medicine andBiology Society. IEEE,2008,2644-2647.
[47] Czosnyka M, Pickard JD. Monitoring and interpretation of intracranial pressure[J]. BritishMedical Journal,2004,75:(6)813.
[48] Mokri B. The Monro–Kellie hypothesis[J]. Neurology,2001,56:(12)1746-1748.
[49] Smith M. Monitoring intracranial pressure in traumatic brain injury[J]. Anesthesia&Analgesia,2008,106:(1)240-248.
[50] Hayashi M, Marukawa S, Fujii H, Kitano T, Kobayashi H, Yamamoto S. Intracranialhypertension in patients with ruptured intracranial aneurysm[J]. Journal of Neurosurgery,1977,46:(5)584-590.
[51] Matsuda M, Yoneda S, Handa H, Gotoh H. Cerebral hemodynamic changes during plateauwaves in brain-tumor patients[J]. Journal of Neurosurgery,1979,50:(4)483-488.
[52] Newell DW, Aaslid R, Stooss R, Reulen HJ. The relationship of blood flow velocityfluctuations to intracranial pressure B waves[J]. Journal of neurosurgery,1992,76:(3)415-421.
[53] Lemaire JJ, Khalil T, Cervenansky F, Gindre G, Boire JY, Bazin JE, Irthum B, Chazal J. Slowpressure waves in the cranial enclosure[J]. Acta neurochirurgica,2002,144:(3)243-254.
[54] Czosnyka M, Czosnyka Z, Keong N, Lavinio A, Smielewski P, Momjian S, Schmidt EA,Petrella G, Owler B, Pickard JD. Pulse pressure waveform in hydrocephalus: what it is andwhat it isn't[J]. Neurosurgical Focus,2007,22:(4)1-7.
[55] Czosnyka M, Smielewski P, Timofeev I, Lavinio A, Guazzo E, Hutchinson P, Pickard JD.Intracranial pressure: more than a number[J]. Neurosurgical Focus,2007,22:(5)1-7.
[56] Czosnyka ZH, Cieslicki K, Czosnyka M, Pickard JD. Hydrocephalus shunts and waves ofintracranial pressure[J]. Medical and Biological Engineering and Computing,2005,43:(1)71-77.
[57] Holm S, Eide PK. The frequency domain versus time domain methods for processing ofintracranial pressure (ICP) signals[J]. Medical engineering&physics,2008,30:(2)164-170.
[58] Eide PK, Egge A, Due-T nnessen BJ, Helseth E. Is Intracranial Pressure Waveform AnalysisUseful in the Management of Pediatric Neurosurgical Patients?[J]. Pediatr Neurosurg,2007,43:(6)472-481.
[59] Scalzo F, Xu P, Asgari S, Bergsneider M, Hu X. Regression analysis for peak designation inpulsatile pressure signals[J]. Medical and Biological Engineering and Computing,2009,47:(9)967-977.
[60] Hu X, Xu P, Scalzo F, Vespa P, Bergsneider M. Morphological clustering and analysis ofcontinuous intracranial pressure[J]. IEEE Transactions on Biomedical Engineering,2009,56:(3)696-705.
[61] Hu X, Xu P, Lee DJ, Vespa P, Baldwin K, Bergsneider M. An algorithm of extractingintracranial pressure latency relative to electrocardiogram r wave[J]. Physiologicalmeasurement,2008,29:(4)459.
[62] Mitchell GF, Pfeffer MA, Finn PV, Pfeffer JM. Comparison of techniques for measuringpulse-wave velocity in the rat[J]. Journal of Applied Physiology,1997,82:(1)203.
[63] Chiu YC, Arand PW, Shroff SG, Feldman T, Carroll JD. Determination of pulse wavevelocities with computerized algorithms[J]. American heart journal,1991,121:(5)1460-1470.
[64] Kazanavicius E, Gircys R, Vrubliauskas A, Lugin S. Mathematical methods for determiningthe foot point of the arterial pulse wave and evaluation of proposed methods[J]. InformationTech Control,2005,34:(1)29-36.
[65] Eide PK. A new method for processing of continuous intracranial pressure signals[J]. Medicalengineering&physics,2006,28:(6)579-587.
[66] Xu P, Bergsneider M, Hu X. Pulse onset detection using neighbor pulse-based signalenhancement[J]. Medical engineering&physics,2009,31:(3)337-345.
[67] Lowe DG. Object recognition from local scale-invariant features[C]. in The Proceedings of theSeventh IEEE International Conference on Computer Vision. IEEE,1999,1150-1157.
[68] Belongie S, Malik J, Puzicha J. Shape matching and object recognition using shape contexts[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,509-522.
[69] Lowe DG. Distinctive image features from scale-invariant keypoints[J]. International journalof computer vision,2004,60:(2)91-110.
[70] Ling H, Jacobs DW. Shape classification using the inner-distance[J]. IEEE transactions onpattern analysis and machine intelligence,2007,29:(2)286.
[71] Srisuk S, Petrou M, Kurutach W, Kadyrov A. A face authentication system using the tracetransform[J]. Pattern Analysis&Applications,2005,8:(1)50-61.
[72] Shechtman E, Irani M. Matching local self-similarities across images and videos[C]. in IEEEConference on Computer Vision and Pattern Recognition. Citeseer,2007,511–518.
[73]刘智勇.基于云计算的文本挖掘算法研究[D],成都：电子科技大学硕士学位论文.2011.
[74] Jain AK, Dubes RX. Algorithms for clustering data[M].1988.
[75] King B. Step-wise clustering procedures[J]. Journal of the American Statistical Association,1967,62:(317)86-101.
[76] Xu P, Bergsneider M, Hu X. Pulse onset detection using neighbor pulse-based signalenhancement[J]. Medical Engineering and Physics,2009,31:(3)337-345.
[77] Dietterich TG. Approximate statistical tests for comparing supervised classification learningalgorithms[J]. Neural computation,1998,10:(7)1895-1923.
[78] Balestreri M, Czosnyka M, Steiner LA, Schmidt E, Smielewski P, Matta B, Pickard JD.Intracranial hypertension: What additional information can be derived from ICP waveformafter head injury?[J]. Acta Neurochirurgica,2004,146:(2)131-141.
[79] Panerai RB. Cerebral autoregulation: from models to clinical applications[J]. CardiovascularEngineering,2008,8:(1)42-59.
[80] Zong W, Heldt T, Moody GB, Mark RG. An open-source algorithm to detect onset of arterialblood pressure pulses[J]. Computers in Cardiology,2003,259-262.
[81] Li BN, Dong MC, Vai MI. On an automatic delineator for arterial blood pressurewaveforms[J]. Biomedical Signal Processing and Control,2010,5:(1)76-81.
[82] Arafat MA, Hasan MK. Automatic detection of ECG wave boundaries using empirical modedecomposition[J].2009.
[83] Martínez JP, Almeida R, Olmos S, Rocha AP, Laguna P. A wavelet-based ECG delineator:evaluation on standard databases[J]. IEEE Transactions on Biomedical Engineering,2004,51:(4)570-581.
[84] Aboy M, McNames J, Thong T, Tsunami D, Ellenby MS, Goldstein B. An automatic beatdetection algorithm for pressure signals[J]. IEEE Transactions on Biomedical Engineering,2005,52:(10)1662-1670.
[85] Hirai O, Handa H, Ishikawa M, Kim SH. Epidural pulse waveform as an indicator ofintracranial pressure dynamics[J]. Surgical neurology,1984,21:(1)75-79.
[86] Scalzo F, Asgari S, Kim S, Bergsneider M, Hu X. Robust peak recognition in intracranialpressure signals[J]. Biomedical engineering online,2010,9:(1)61.
[87] Grenander U, Rosenblatt M. Statistical analysis of stationary time series[M]: Wiley New York.1957.
[88] Blumer A, Ehrenfeucht A, Haussler D, Warmuth MK. Learnability and theVapnik-Chervonenkis dimension[J]. Journal of the ACM (JACM),1989,36:(4)929-965.
[89] Langley P. Elements of machine learning[M]: Morgan Kaufmann Pub.1996.
[90] Ashley KD, Brüninghaus S. Automatically classifying case texts and predicting outcomes[J].Artificial Intelligence and Law,2009,17:(2)125-165.
[91] Alpaydin E. Introduction to machine learning[M]: The MIT Press.2004.
[92] Kimeldorf G, Wahba G. Some results on Tchebycheffian spline functions[J]. Journal ofMathematical Analysis and Applications,1971,33:(1)82-95.
[93] Boser BE, Guyon IM, Vapnik VN. A training algorithm for optimal margin classifiers[C].ACM,1992,144-152.
[94] Vapnik VN, Vapnik V. Statistical learning theory[M], vol.2: Wiley New York.1998.
[95] Kaufman L. Solving the Quadratic Programming Problem Arising in Support VectorClassification[M], in: Advances in Kernel Methods. MIT Press, Cambridge, MA, USA.1999,pp147-168.
[96] Vapnik VN. The nature of statistical learning theory[M]: Springer Verlag.2000.
[97] Cortes C, Vapnik V. Support-vector networks[J]. Machine learning,1995,20:(3)273-297.
[98] Selvaraj H, Selvi ST, Selvathi D, Gewali L. Brain MRI slices classification using least squaressupport vector machine[J]. International Journal of Intelligent Computing in Medical Sciencesand Image Processing,2007,1:(1)21-33.
[99]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26:(1)32-42.
[100] Rifkin R, Klautau A. In defense of one-vs-all classification[J]. The Journal of MachineLearning Research,2004,5:101-141.
[101] Chang CC, Lin CJ. LIBSVM: a library for support vector machines[EB/01].2001, Availablefrom: www.csie.ntu.edu.tw/~cjlin/libsvm/.
[102] Yang L, Zhao M, Peng C, Hu X, Feng H, Ji Z. Waveform descriptor for pulse onset detectionof intracranial pressure signal[J]. Medical Engineering&Physics,2012,34:179-186.
[103] Kohavi R. A study of cross-validation and bootstrap for accuracy estimation and modelselection[C]. LAWRENCE ERLBAUM ASSOCIATES LTD,1995,1137-1145.
[104] Portet F. P wave detector with PP rhythm tracking: evaluation in different arrhythmiacontexts[J]. Physiological measurement,2008,29:141.
[105] Storn R, Price K. Differential evolution-a simple and efficient adaptive scheme for globaloptimization over continuous spaces[R]. Technical Report Technical Report TR-95-012.International Computer Science Institute.1995. http://http.icsi.berkeley.edu/-storn/litera.html.
[106] Storn R, Price K. Differential evolution–a simple and efficient heuristic for globaloptimization over continuous spaces[J]. Journal of global optimization,1997,11:(4)341-359.
[107]刘波,王凌,金以慧.差分进化算法研究进展[J].控制与决策,2007,22:(7)721-729.
[108]袁晓辉,苏安俊,聂浩,张勇传,袁艳斌.差分进化算法在电力系统中的应用研究进展[J].华东电力,2009,37:(002)243-249.
[109] G mperle R, Müller SD, Koumoutsakos P. A parameter study for differential evolution[J].Advances in Intelligent Systems, Fuzzy Systems, Evolutionary Computation,2002,10:293-298.
[110]陈友,程学旗,李洋,戴磊.基于特征选择的轻量级入侵检测系统[J].软件学报,2007,18:(7)1639-1651.
[111] Dash M, Liu H. Feature selection for classification[J]. Intelligent data analysis,1997,1:(1-4)131-156.
[112] Gutierrez-Osuna R.[EB/01].2000, Available from: http://research.cs.tamu.edu/prism/lectures/talks/nose00.pdf.
[113] Saeys Y, Inza I, Larra aga P. A review of feature selection techniques in bioinformatics[J].Bioinformatics,2007,23:(19)2507.
[114]张齐,林媛媛,余国先.一种基于图的特征选择方法[J]. Computer Engineering andApplications,2011,47:(26).
[115] Hu X, Nenov V, Bergsneider M, Glenn TC, Vespa P, Martin N. Estimation of hidden statevariables of the intracranial system using constrained nonlinear Kalman filters[J]. IEEETransactions on Biomedical Engineering,2007,54:(4)597-610.
[116] Hu X, Glenn T, Scalzo F, Bergsneider M, Sarkiss C, Martin N, Vespa P. Intracranial pressurepulse morphological features improved detection of decreased cerebral blood flow[J].Physiological measurement,2010,31:679.
[117] B rgesen SE, Gjerris F, S rensen SC. Intracranial pressure and conductance to outflow ofcerebrospinal fluid in normal-pressure hydrocephalus[J]. Journal of neurosurgery,1979,50:(4)489-493.
[118] Ursino M, Di Giammarco P. A mathematical model of the relationship between cerebral bloodvolume and intracranial pressure changes: the generation of plateau waves[J]. Annals ofbiomedical engineering,1991,19:(1)15-42.
[119] Nedelmann M, Kaps M, Mueller-Forell W. Venous obstruction and jugular valveinsufficiency in idiopathic intracranial hypertension[J]. Journal of neurology,2009,256:(6)964-969.
[120] Stevens SA, Stimpson J, Lakin WD, Thakore NJ, Penar PL. A model for idiopathicintracranial hypertension and associated pathological ICP wave-forms[J]. IEEE Transactionson Biomedical Engineering,2008,55:(2)388-398.
[121] Marmarou A, Shulman K, Rosende RM. A nonlinear analysis of the cerebrospinal fluidsystem and intracranial pressure dynamics[J]. Journal of Neurosurgery,1978,48:(3)332-344.
[122] Zhu X. Semi-supervised learning literature survey[R]. Technical Report TR-1530. ComputerScience, University of Wisconsin-Madison.2006.
[123] Carlson A, Betteridge J, Wang RC, Hruschka Jr ER, Mitchell TM. Coupled semi-supervisedlearning for information extraction[C]. ACM,2010,101-110.
[124] Cao L. Support vector machines experts for time series forecasting[J]. Neurocomputing,2003,51:321-339
[2]岳献芳,王立,周峰.颅内压变化引起的颅骨表面应变分析[J].北京科技大学学报,2006,28:(012)1143-1151.
[3] Jennum P, B rgesen SE. Intracranial pressure and obstructive sleep apnea[J]. Chest,1989,95:(2)279-283.
[4] Marmarou A, Anderson RL, Ward JD, Choi SC, Young HF, Eisenberg HM, Foulkes MA,Marshall LF, Jane JA. Impact of ICP instability and hypotension on outcome in patients withsevere head trauma[J]. Special Supplements,1991,75:(1S)59-66.
[5]钱素云,樊寻梅.急性颅内压增高与应激反应[J]. PEDIATRIC EMERGENCY MEDICINE,2003,10:(2).
[6]王群思.颅内压监测在危重症监护中的应用[J]. CHINESE PEDIATRIC EMERGENCYMEDICINE,2007,14:(3).
[7]李振举.颅脑损伤后脑血流变化及自由基反应对预后的影响[D],长春：吉林大学硕士学位论文.2007.
[8] Ohta H, Hadeishi H, Nemoto M, Kawamura S, Hinuma Y, Suzuki E. Transient Effect ofHyperbaric Oxygen on Cerebral Blood Flow and Intracranial Pressure[J].1990.
[9]杨勇,杨义,王任直.颅内压监测技术的进展[J].中国微侵袭神经外科杂志,2008,13:(6)284.
[10]苏治国,张赛.高渗盐水对脑水肿和颅内高压治疗作用的研究进展[J].武警医学院学报,2006,15:(006)607-610.
[11]徐姜定.颅内压增高监测及治疗进展[J].浙江临床医学,2003,5:(7)481-482.
[12] Guttag JJ, Ganeshapillai G. Methods to improve the signal quality of corruptedmulti-parameter physiological signals[D], Massachusetts Institute of Technology.2011.
[13] Hornero R, Aboy M, Abásolo D, McNames J, Goldstein B. Interpretation of approximateentropy: analysis of intracranial pressure approximate entropy during acute intracranialhypertension[J]. IEEE Transactions on Biomedical Engineering,2005,52:(10)1671-1680.
[14] Hamilton R, Xu P, Asgari S, Kasprowicz M, Vespa P, Bergsneider M, Hu X. Forecastingintracranial pressure elevation using pulse waveform morphology[C]. IEEE,2009,4331-4334.
[15] Wiklund L, Hok B, Stahl K, Jordeby-Jonsson A. Postanesthesia monitoring revisited:frequency of true and false alarms from different monitoring devices[J]. Journal of clinicalanesthesia,1994,6:(3)182-188.
[16] Lawless ST. Crying wolf: false alarms in a pediatric intensive care unit[J]. Critical caremedicine,1994,22:(6)981.
[17] Imhoff M, Bauer M, Gather U, Fried R. Pattern detection in intensive care monitoring timeseries with autoregressive models: Influence of the model order[J]. Biometrical Journal,2002,44:(6)746-761.
[18] Ursino M, Iezzi M, Stocchetti N. Intracranial pressure dynamics in patients with acute braindamage: a critical analysis with the aid of a mathematical model[J]. IEEE Transactions onBiomedical Engineering,1995,42:(6)529-540.
[19] Swiercz M, Mariak Z, Lewko J, Chojnacki K, Kozlowski A, Piekarski P. Neural networktechnique for detecting emergency states in neurosurgical patients[J]. Medical and BiologicalEngineering and Computing,1998,36:(6)717-722.
[20] Charbonnier S, Becq G, Biot L. On-line segmentation algorithm for continuously monitoreddata in intensive care units[J]. IEEE Transactions on Biomedical Engineering,2004,51:(3)484-492
[21] Charbonnier S, Gentil S. A trend-based alarm system to improve patient monitoring inintensive care units[J]. Control engineering practice,2007,15:(9)1039-1050.
[22] Hinkley DV. Inference about the change-point from cumulative sum tests[J]. Biometrika,1971,58:(3)509.
[23] Keogh E, Chu S, Hart D, Pazzani M. An online algorithm for segmenting time series[C]. inProceedings IEEE International Conference on Data Mining. IEEE,2001,289-296.
[24] Kirkness CJ, Mitchell PH, Burr RL, March KS, Newell DW. Intracranial pressure waveformanalysis: clinical and research implications[J]. The Journal of neuroscience nursing: journal ofthe American Association of Neuroscience Nurses,2000,32:(5)271.
[25] Swiercz M, Mariak Z, Krejza J, Lewko J, Szydlik P. Intracranial pressure processing withartificial neural networks: Prediction of ICP trends[J]. Acta neurochirurgica,2000,142:(4)401-406.
[26] Fan JY, Kirkness C, Vicini P, Burr R, Mitchell P. Intracranial Pressure WaveformMorphology and Intracranial Adaptive Capacity[J]. American Journal of Critical Care,2008,17:(6)545.
[27] Westhout FD, Paré LS, Delfino RJ, Cramer SC. Slope of the intracranial pressure waveformafter traumatic brain injury[J]. Surgical neurology,2008,70:(1)70-74.
[28] Hickey JV, Olson DWM, Turner DA. Intracranial pressure waveform analysis during rest andsuctioning[J]. Biological Research For Nursing,2009,11:(2)174.
[29] Contant CF. Intracranial pressure waveform indices in transient and refractory intracranialhypertension[J]. Journal of Neuroscience methods,1995,57:(1)15-25.
[30] Lundberg N. Continuous recording and control of ventricular fluid pressure in neurosurgicalpractice[J]. Acta psychiatrica Scandinavica. Supplementum,1960,36:(149)1.
[31] Czosnyka M, Smielewski P, Piechnik S, Schmidt EA, Al-Rawi PG, Kirkpatrick PJ, Pickard JD.Hemodynamic characterization of intracranial pressure plateau waves in head-injuredpatients[J]. Journal of neurosurgery,1999,91:(1)11-19.
[32] Szewczykowski J, Dytko P, Kunicki A, Korsak-Sliwka J, Sliwka S, Dziduszko J, AugustyniakB. Determination of critical ICP levels in neurosurgical patients: a statistical approach[J].Intracranial Pressure II,1975,392-393.
[33] Turner JM, McDowall DG, Gibson RM, Khaili H. Computer analysis of intracranial pressuremeasurements: Clinical value and nursing response[J]. Intracranial pressure III,1976,283-287.
[34] Price DJ, Dugdale RE, Mason J. The control of ICP using three asynchronous closed loops[J].Intracranial Pressure IV,1980,395-399.
[35] Portnoy HD, Chopp M. Cerebrospinal fluid pulse wave form analysis during hypercapnia andhypoxia[J]. Neurosurgery,1981,9:(1)14.
[36] McNames J, Crespo C, Bassale J, Aboy M, Elleby M, Lai S, Goldstein B. Sensitive precursorsto acute episodes of intracranial hypertension[C]. in The4th International Workshop inBiosignal Interpretation. Como, Italy,2002,303-306.
[37] Pincus SM. Approximate entropy as a measure of system complexity[J]. Proceedings of theNational Academy of Sciences,1991,88:(6)2297.
[38] Hornero R, Aboy M, Abasolo D, McNames J, Wakeland W, Goldstein B. Complex analysis ofintracranial hypertension using approximate entropy[J]. Critical care medicine,2006,34:(1)87.
[39] Hu X, Miller C, Vespa P, Bergsneider M. Adaptive computation of approximate entropy andits application in integrative analysis of irregularity of heart rate variability and intracranialpressure signals[J]. Medical engineering&physics,2008,30:(5)631-639.
[40] Hu X, Xu P, Asgari S, Vespa P, Bergsneider M. Forecasting ICP elevation based on prescientchanges of intracranial pressure waveform morphology[J]. IEEE Transactions on BiomedicalEngineering,2010,57:(5)1070-1078.
[41] Eide PK, Czosnyka M, Sorteberg W, Pickard JD, Smielewski P. Association betweenintracranial, arterial pulse pressure amplitudes and cerebral autoregulation in head injurypatients[J]. Neurological Research,2007,29:(6)578-582.
[42]姚旭,王晓丹,张玉玺,权文.特征选择方法综述[J].控制与决策,2012.
[43] Weston J, Mukherjee S, Chapelle O, Pontil M, Poggio T, Vapnik V. Feature selection forSVMs[J]. Advances in neural information processing systems,2001,668-674.
[44] Maxeiner H, Behnke M. Intracranial volume, brain volume, reserve volume andmorphological signs of increased intracranial pressure-A post-mortem analysis[J]. LegalMedicine,2008,10:(6)293-300.
[45]刘宏雨.侧脑室引流在68例脑室出血患者中的应用及护理[J].中华护理杂志,2005,40:(007)516-518.
[46] Shahsavari S, McKelvey T. Harmonics tracking of intracranial and arterial blood pressurewaves[C]. in30th Annual IEEE International Conference on Engineering in Medicine andBiology Society. IEEE,2008,2644-2647.
[47] Czosnyka M, Pickard JD. Monitoring and interpretation of intracranial pressure[J]. BritishMedical Journal,2004,75:(6)813.
[48] Mokri B. The Monro–Kellie hypothesis[J]. Neurology,2001,56:(12)1746-1748.
[49] Smith M. Monitoring intracranial pressure in traumatic brain injury[J]. Anesthesia&Analgesia,2008,106:(1)240-248.
[50] Hayashi M, Marukawa S, Fujii H, Kitano T, Kobayashi H, Yamamoto S. Intracranialhypertension in patients with ruptured intracranial aneurysm[J]. Journal of Neurosurgery,1977,46:(5)584-590.
[51] Matsuda M, Yoneda S, Handa H, Gotoh H. Cerebral hemodynamic changes during plateauwaves in brain-tumor patients[J]. Journal of Neurosurgery,1979,50:(4)483-488.
[52] Newell DW, Aaslid R, Stooss R, Reulen HJ. The relationship of blood flow velocityfluctuations to intracranial pressure B waves[J]. Journal of neurosurgery,1992,76:(3)415-421.
[53] Lemaire JJ, Khalil T, Cervenansky F, Gindre G, Boire JY, Bazin JE, Irthum B, Chazal J. Slowpressure waves in the cranial enclosure[J]. Acta neurochirurgica,2002,144:(3)243-254.
[54] Czosnyka M, Czosnyka Z, Keong N, Lavinio A, Smielewski P, Momjian S, Schmidt EA,Petrella G, Owler B, Pickard JD. Pulse pressure waveform in hydrocephalus: what it is andwhat it isn't[J]. Neurosurgical Focus,2007,22:(4)1-7.
[55] Czosnyka M, Smielewski P, Timofeev I, Lavinio A, Guazzo E, Hutchinson P, Pickard JD.Intracranial pressure: more than a number[J]. Neurosurgical Focus,2007,22:(5)1-7.
[56] Czosnyka ZH, Cieslicki K, Czosnyka M, Pickard JD. Hydrocephalus shunts and waves ofintracranial pressure[J]. Medical and Biological Engineering and Computing,2005,43:(1)71-77.
[57] Holm S, Eide PK. The frequency domain versus time domain methods for processing ofintracranial pressure (ICP) signals[J]. Medical engineering&physics,2008,30:(2)164-170.
[58] Eide PK, Egge A, Due-T nnessen BJ, Helseth E. Is Intracranial Pressure Waveform AnalysisUseful in the Management of Pediatric Neurosurgical Patients?[J]. Pediatr Neurosurg,2007,43:(6)472-481.
[59] Scalzo F, Xu P, Asgari S, Bergsneider M, Hu X. Regression analysis for peak designation inpulsatile pressure signals[J]. Medical and Biological Engineering and Computing,2009,47:(9)967-977.
[60] Hu X, Xu P, Scalzo F, Vespa P, Bergsneider M. Morphological clustering and analysis ofcontinuous intracranial pressure[J]. IEEE Transactions on Biomedical Engineering,2009,56:(3)696-705.
[61] Hu X, Xu P, Lee DJ, Vespa P, Baldwin K, Bergsneider M. An algorithm of extractingintracranial pressure latency relative to electrocardiogram r wave[J]. Physiologicalmeasurement,2008,29:(4)459.
[62] Mitchell GF, Pfeffer MA, Finn PV, Pfeffer JM. Comparison of techniques for measuringpulse-wave velocity in the rat[J]. Journal of Applied Physiology,1997,82:(1)203.
[63] Chiu YC, Arand PW, Shroff SG, Feldman T, Carroll JD. Determination of pulse wavevelocities with computerized algorithms[J]. American heart journal,1991,121:(5)1460-1470.
[64] Kazanavicius E, Gircys R, Vrubliauskas A, Lugin S. Mathematical methods for determiningthe foot point of the arterial pulse wave and evaluation of proposed methods[J]. InformationTech Control,2005,34:(1)29-36.
[65] Eide PK. A new method for processing of continuous intracranial pressure signals[J]. Medicalengineering&physics,2006,28:(6)579-587.
[66] Xu P, Bergsneider M, Hu X. Pulse onset detection using neighbor pulse-based signalenhancement[J]. Medical engineering&physics,2009,31:(3)337-345.
[67] Lowe DG. Object recognition from local scale-invariant features[C]. in The Proceedings of theSeventh IEEE International Conference on Computer Vision. IEEE,1999,1150-1157.
[68] Belongie S, Malik J, Puzicha J. Shape matching and object recognition using shape contexts[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,509-522.
[69] Lowe DG. Distinctive image features from scale-invariant keypoints[J]. International journalof computer vision,2004,60:(2)91-110.
[70] Ling H, Jacobs DW. Shape classification using the inner-distance[J]. IEEE transactions onpattern analysis and machine intelligence,2007,29:(2)286.
[71] Srisuk S, Petrou M, Kurutach W, Kadyrov A. A face authentication system using the tracetransform[J]. Pattern Analysis&Applications,2005,8:(1)50-61.
[72] Shechtman E, Irani M. Matching local self-similarities across images and videos[C]. in IEEEConference on Computer Vision and Pattern Recognition. Citeseer,2007,511–518.
[73]刘智勇.基于云计算的文本挖掘算法研究[D],成都：电子科技大学硕士学位论文.2011.
[74] Jain AK, Dubes RX. Algorithms for clustering data[M].1988.
[75] King B. Step-wise clustering procedures[J]. Journal of the American Statistical Association,1967,62:(317)86-101.
[76] Xu P, Bergsneider M, Hu X. Pulse onset detection using neighbor pulse-based signalenhancement[J]. Medical Engineering and Physics,2009,31:(3)337-345.
[77] Dietterich TG. Approximate statistical tests for comparing supervised classification learningalgorithms[J]. Neural computation,1998,10:(7)1895-1923.
[78] Balestreri M, Czosnyka M, Steiner LA, Schmidt E, Smielewski P, Matta B, Pickard JD.Intracranial hypertension: What additional information can be derived from ICP waveformafter head injury?[J]. Acta Neurochirurgica,2004,146:(2)131-141.
[79] Panerai RB. Cerebral autoregulation: from models to clinical applications[J]. CardiovascularEngineering,2008,8:(1)42-59.
[80] Zong W, Heldt T, Moody GB, Mark RG. An open-source algorithm to detect onset of arterialblood pressure pulses[J]. Computers in Cardiology,2003,259-262.
[81] Li BN, Dong MC, Vai MI. On an automatic delineator for arterial blood pressurewaveforms[J]. Biomedical Signal Processing and Control,2010,5:(1)76-81.
[82] Arafat MA, Hasan MK. Automatic detection of ECG wave boundaries using empirical modedecomposition[J].2009.
[83] Martínez JP, Almeida R, Olmos S, Rocha AP, Laguna P. A wavelet-based ECG delineator:evaluation on standard databases[J]. IEEE Transactions on Biomedical Engineering,2004,51:(4)570-581.
[84] Aboy M, McNames J, Thong T, Tsunami D, Ellenby MS, Goldstein B. An automatic beatdetection algorithm for pressure signals[J]. IEEE Transactions on Biomedical Engineering,2005,52:(10)1662-1670.
[85] Hirai O, Handa H, Ishikawa M, Kim SH. Epidural pulse waveform as an indicator ofintracranial pressure dynamics[J]. Surgical neurology,1984,21:(1)75-79.
[86] Scalzo F, Asgari S, Kim S, Bergsneider M, Hu X. Robust peak recognition in intracranialpressure signals[J]. Biomedical engineering online,2010,9:(1)61.
[87] Grenander U, Rosenblatt M. Statistical analysis of stationary time series[M]: Wiley New York.1957.
[88] Blumer A, Ehrenfeucht A, Haussler D, Warmuth MK. Learnability and theVapnik-Chervonenkis dimension[J]. Journal of the ACM (JACM),1989,36:(4)929-965.
[89] Langley P. Elements of machine learning[M]: Morgan Kaufmann Pub.1996.
[90] Ashley KD, Brüninghaus S. Automatically classifying case texts and predicting outcomes[J].Artificial Intelligence and Law,2009,17:(2)125-165.
[91] Alpaydin E. Introduction to machine learning[M]: The MIT Press.2004.
[92] Kimeldorf G, Wahba G. Some results on Tchebycheffian spline functions[J]. Journal ofMathematical Analysis and Applications,1971,33:(1)82-95.
[93] Boser BE, Guyon IM, Vapnik VN. A training algorithm for optimal margin classifiers[C].ACM,1992,144-152.
[94] Vapnik VN, Vapnik V. Statistical learning theory[M], vol.2: Wiley New York.1998.
[95] Kaufman L. Solving the Quadratic Programming Problem Arising in Support VectorClassification[M], in: Advances in Kernel Methods. MIT Press, Cambridge, MA, USA.1999,pp147-168.
[96] Vapnik VN. The nature of statistical learning theory[M]: Springer Verlag.2000.
[97] Cortes C, Vapnik V. Support-vector networks[J]. Machine learning,1995,20:(3)273-297.
[98] Selvaraj H, Selvi ST, Selvathi D, Gewali L. Brain MRI slices classification using least squaressupport vector machine[J]. International Journal of Intelligent Computing in Medical Sciencesand Image Processing,2007,1:(1)21-33.
[99]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26:(1)32-42.
[100] Rifkin R, Klautau A. In defense of one-vs-all classification[J]. The Journal of MachineLearning Research,2004,5:101-141.
[101] Chang CC, Lin CJ. LIBSVM: a library for support vector machines[EB/01].2001, Availablefrom: www.csie.ntu.edu.tw/~cjlin/libsvm/.
[102] Yang L, Zhao M, Peng C, Hu X, Feng H, Ji Z. Waveform descriptor for pulse onset detectionof intracranial pressure signal[J]. Medical Engineering&Physics,2012,34:179-186.
[103] Kohavi R. A study of cross-validation and bootstrap for accuracy estimation and modelselection[C]. LAWRENCE ERLBAUM ASSOCIATES LTD,1995,1137-1145.
[104] Portet F. P wave detector with PP rhythm tracking: evaluation in different arrhythmiacontexts[J]. Physiological measurement,2008,29:141.
[105] Storn R, Price K. Differential evolution-a simple and efficient adaptive scheme for globaloptimization over continuous spaces[R]. Technical Report Technical Report TR-95-012.International Computer Science Institute.1995. http://http.icsi.berkeley.edu/-storn/litera.html.
[106] Storn R, Price K. Differential evolution–a simple and efficient heuristic for globaloptimization over continuous spaces[J]. Journal of global optimization,1997,11:(4)341-359.
[107]刘波,王凌,金以慧.差分进化算法研究进展[J].控制与决策,2007,22:(7)721-729.
[108]袁晓辉,苏安俊,聂浩,张勇传,袁艳斌.差分进化算法在电力系统中的应用研究进展[J].华东电力,2009,37:(002)243-249.
[109] G mperle R, Müller SD, Koumoutsakos P. A parameter study for differential evolution[J].Advances in Intelligent Systems, Fuzzy Systems, Evolutionary Computation,2002,10:293-298.
[110]陈友,程学旗,李洋,戴磊.基于特征选择的轻量级入侵检测系统[J].软件学报,2007,18:(7)1639-1651.
[111] Dash M, Liu H. Feature selection for classification[J]. Intelligent data analysis,1997,1:(1-4)131-156.
[112] Gutierrez-Osuna R.[EB/01].2000, Available from: http://research.cs.tamu.edu/prism/lectures/talks/nose00.pdf.
[113] Saeys Y, Inza I, Larra aga P. A review of feature selection techniques in bioinformatics[J].Bioinformatics,2007,23:(19)2507.
[114]张齐,林媛媛,余国先.一种基于图的特征选择方法[J]. Computer Engineering andApplications,2011,47:(26).
[115] Hu X, Nenov V, Bergsneider M, Glenn TC, Vespa P, Martin N. Estimation of hidden statevariables of the intracranial system using constrained nonlinear Kalman filters[J]. IEEETransactions on Biomedical Engineering,2007,54:(4)597-610.
[116] Hu X, Glenn T, Scalzo F, Bergsneider M, Sarkiss C, Martin N, Vespa P. Intracranial pressurepulse morphological features improved detection of decreased cerebral blood flow[J].Physiological measurement,2010,31:679.
[117] B rgesen SE, Gjerris F, S rensen SC. Intracranial pressure and conductance to outflow ofcerebrospinal fluid in normal-pressure hydrocephalus[J]. Journal of neurosurgery,1979,50:(4)489-493.
[118] Ursino M, Di Giammarco P. A mathematical model of the relationship between cerebral bloodvolume and intracranial pressure changes: the generation of plateau waves[J]. Annals ofbiomedical engineering,1991,19:(1)15-42.
[119] Nedelmann M, Kaps M, Mueller-Forell W. Venous obstruction and jugular valveinsufficiency in idiopathic intracranial hypertension[J]. Journal of neurology,2009,256:(6)964-969.
[120] Stevens SA, Stimpson J, Lakin WD, Thakore NJ, Penar PL. A model for idiopathicintracranial hypertension and associated pathological ICP wave-forms[J]. IEEE Transactionson Biomedical Engineering,2008,55:(2)388-398.
[121] Marmarou A, Shulman K, Rosende RM. A nonlinear analysis of the cerebrospinal fluidsystem and intracranial pressure dynamics[J]. Journal of Neurosurgery,1978,48:(3)332-344.
[122] Zhu X. Semi-supervised learning literature survey[R]. Technical Report TR-1530. ComputerScience, University of Wisconsin-Madison.2006.
[123] Carlson A, Betteridge J, Wang RC, Hruschka Jr ER, Mitchell TM. Coupled semi-supervisedlearning for information extraction[C]. ACM,2010,101-110.
[124] Cao L. Support vector machines experts for time series forecasting[J]. Neurocomputing,2003,51:321-339