The Study of Probing Patterns for Physiological Signals Using Empirical Markov Chain Matrix

• Main Authors: LAN,YI-BIN, 藍翊賓
• Other Authors: Joesph Arul
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/vguxmq

碩士 === 輔仁大學 === 資訊工程學系碩士班 === 107 === Abstract A physiological activity is often referred to a type of signal in time series. The purpose of this thesis was to use the data collected from the heartbeat intensity of shrimps under Electro Cardio Gram (ECG) to find their patterns of physiological signals, to provide further details of the behavior of shrimps. The physiological signals are classified into two categories: biologically significant waveforms and the differences between the given times of peaks as targets for detection. The detection method consists of the principal analysis and ķ-means clustering, and applying Markov chain matrix method. When the signals are recorded in sequential events, it is significant to probe the patterns which represent a certain physiological or pathological situation in animal behavior. The ECG is a very important clinical diagnostic tool. The various parameters derived from it are not only useful for the diagnosis of cardiovascular diseases, but also can be used as indicators of autonomic nervous activity. To probe the pattern of ECG, in this thesis a new method is proposed with two steps: The first step is to determine an empirical Markov Chain matrix, and the second step is to find possible candidates of patterns from the Markov Chain matrix and classify the candidates as patterns. It is applied to the physiological signal of the time difference between the waveform and the peak. Keywords: physiological signal, pattern, Markov chain matrix