论文部分内容阅读
准确区分生理序列随机性与混沌性,且不受序列长度与参数的影响是衡量复杂度算法的关键。本文提出了一种编码式Lempel-Ziv(LZ)算法,分别从序列随机性与混沌性的区分、长度的影响、动力学性质突变的敏感性、高斯白与粉红噪声复杂度测量等4个方面与经典LZ算法、多状态LZ算法、样本熵以及排列熵进行比较。结果显示,在短、中、长时(100、500、5 000点)下,编码式LZ算法均能准确区分随机与混沌性,正确测度高斯噪声的复杂度低于粉红噪声,并能准确响应序列动力学性质的改变。本文采用美国麻省理工学院(MIT)和波士顿贝斯以色列医院(BIH)联合建立的的MIT-BIH心电数据库中的充血性心力衰竭RR间期(CHF-RR)数据和正常窦性心律RR间期(NSR-RR)数据进行测试,实验结果显示,在各种时长下,编码式LZ复杂度算法均能准确地得出心力衰竭的复杂度低于窦性心律(P<0.01)的结果,且不受长度与参数影响,具有较强的泛化能力。
It is the key to measure the complexity algorithm to accurately distinguish the randomness and chaos of the physiological sequence and not affect the sequence length and parameters. In this paper, an encoding Lempel-Ziv (LZ) algorithm is proposed. From the four aspects of the distinction between randomness and chaos, the influence of length, the sensitivity of sudden change of dynamic properties, Gaussian white and pink noise complexity measurement, Compared with classical LZ algorithm, multi-state LZ algorithm, sample entropy and permutation entropy. The results show that the coding LZ algorithm can accurately distinguish randomness and chaos in the short, medium and long time (100, 500, 5000 points), and the Gaussian noise is less complex than the pink noise in the correct measure and can accurately respond Changes in the sequence kinetic properties. In this paper, the RR-RR data of congestive heart failure (CHF-RR) in the MIT-BIH ECG database jointly established by Massachusetts Institute of Technology (MIT) and Boston Beth Israel Hospital (BIH) (NSR-RR) data were tested. The experimental results show that the encoding LZ complexity algorithm can accurately draw the conclusion that the complexity of heart failure is lower than that of sinus rhythm (P <0.01) under various time slots, And not affected by the length and parameters, with a strong generalization ability.