The Theory of Approximate Entropy and its Application / 中国医学物理学杂志

ABSTRACT

Objective: To introduce the advantages of approximate entropy (ApEn) in analyzing biological signals, to discuss the influences of time series' parameters on ApEn and elaborate the present conditions and prospects of its application on medicine. Methods: According to the definition and algorithm, ApEn has advantages on estimating the complexity of signal comparing with other non-linear analysis methods such as correlation dimension, fractal dimension, lyapunov index etc. Based on ApEn algorithm, we formulate programme to quantificationlly analyze the relationship between ApEn and each parameter of time series; Human EEG and ECG and other biological signals can reflect the state of the body. Through detecting changes of biological signals'ApEn, we can detect and monitor the state of an organization. Results: ApEn has many advantages such as a good anti-noise ability, the short date and so on. It makes up the defects of the correlation dimension ect. ApEn has nothing with the amplitude of time series. In the range of low-frequency, ApEn almost increases with frequency. When the date length is more than 1000 points, ApEn is stable. It quantificationlly describes that ApEn only requires short date. Recently. ApEn is widely used in diagnostics, monitoring, anesthesiology, and achieves good results. Conclusion: ApEn is a new developed non-linear analysis method which can estimate the complexity of the signal quantificationlly. It provides a new approach for diagnosing and monitoring a number of diseases. It has a wide application prospect.