本論文主旨在開發睡眠狀態辨識演算法，用以辨識一般人整晚的睡眠狀態。演算法ㄧ開始先將整晚的睡眠脈搏訊號做訊號前處理及雜訊消除，並將每個視窗的脈搏訊號依照時域分析、頻域分析和非線性分析三大種類計算出共32個特徵，然後將每個特徵正規化以消除生理訊號之特徵值單位不同。接著利用隨機減少多數法(random under-sampling)的方法來處理資料不平衡的問題，平衡後的資料再進行特徵降維。本論文以主成分分析(principal component analysis, PCA)及線性判別分析(linear discriminant analysis, LDA)兩種不同降維方法做比較，降維後的特徵作為分類器的輸入參數。此外，本論文使用了三種分類器來做睡眠狀態辨識的比較，第一種使用最常見且方便的最近鄰居法(k-nearest neighbors, KNN)，做為比較依據。第二種分類器使用具時間及隨機特性的分類器隱藏馬可夫模型(hidden Markov model, HMM)，配合具時間序列的睡眠資料來做分析。第三種分類器使用遞迴類神經網路(recurrent neural network, RNN)，此分類器具有學習能力和時間特性兩大優點。本論文利用多導睡眠儀(polysomnography, PSG)收錄了34位健康受試者的資料，並將睡眠狀態分成清醒(wake)、快速動眼期(rapid eye movement, REM)和非快速動眼期(non-rapid eye movement, NREM)三個時期，最近鄰居法平均正確率(accuracy)為80.18%，隱藏馬可夫模型平均正確率為82.65%，遞迴類神經網路平均正確率為83.79%。實驗結果成功的驗證了使用脈搏訊號辨識睡眠狀態之有效性，且在未來可進一步實現在穿戴式載具上，讓睡眠監測更貼近生活且更加的便利。

This thesis proposes an algorithm for sleep stage recognition using pulse rate variability analysis. The algorithm starts with signal preprocessing, that is, signals are preprocessed in order to remove artifacts from the pulse rate signals collected by sensors. Then, with the preprocessing signal, 32 features are generated from the time-domain, frequency-domain and nonlinear analysis. The features are normalized in order to minimize the effect of differences in the ranges of values among different features. Subsequently, the random under-sampling method is utilized to remove data imbalance. This thesis applied principal component analysis (PCA) and linear discriminant analysis (LDA) to do feature dimension reduction and performance comparison. In the sleep stage recognition, three classifiers were applied and compared: a k-nearest neighbors (KNN), a hidden Markov model (HMM), and a recurrent neural network (RNN). A database with classified sleep stage wake, rapid eye movement (REM) and non-rapid eye movement (NREM) of 34 healthy subjects obtained from a hospital in the southern Taiwan was used in the experiment. The average accuracy of KNN, HMM, and RNN were 80.18%, 82.65% and 83.79%, respectively. The effectiveness of the proposed algorithm has validated by the experimental results. In the future, the proposed algorithm can be applied to wearable devices for home sleep monitoring.

[1]H. W. J. Agnew, W. B. Webb, and R. L. Williams, “The first night effect: An EEG study of sleep,” Psychophysiology, vol. 2, no. 3, pp. 263-266, 1966.
[2]M. Adnane, Z. Jiang, and Z. Yan, “Sleep–wake stages classification and sleep efficiency estimation using single-lead electrocardiogram,” Expert Syst. Appl., vol. 39, pp. 1401–1413, 2012.
[3]P. Bušek, J. Vaòková, J. Opavský, J. Salinger, and S. Nevšímalová, “Spectral Analysis of Heart Rate Variability in Sleep”, Physiol. Res, vol. 54, pp. 369-376, 2005.
[4]H. Cruse, Neural Networks as Cybernetic Systems, Brains, Minds & Media, vol. 2, 2006.
[5]H. Danker-Hopfe, D. Kunz, G. Gruber, G. Kloesch, J. L. Lorenzo, S. L. Himanen, B. Kemp, T. Penzel, J. Roschke, H. Dorn, A. Schloegl, E. Trenker, and G. Dorffner, “Interrater reliability between scorers from eight European sleep laboratories in subjects with different sleep disorders,” J. Sleep Res., vol. 13, no. 1, pp. 63-69, 2004.
[6]S. R. Eddy, “Hidden Markov models,” Current Opinion in Structural Biology, vol. 6, pp. 361–365, 1996.
[7]L. J. Epstein, D. Kristo, P. J. Jr. Strollo, et al., “Clinical guideline for the evaluation, management and long-term care of obstructive sleep apnea in adults,” J. of Clinical Sleep Medicine, vol. 5, no. 3, pp. 263-276, 2009.
[8]S. Güneş, K. Polat, and Ş. Yosunkaya, “Efficient sleep stage recognition system based on EEG signal using k-means clustering based feature weighting,” Expert Systems with Applications, vol. 37, no. 12, pp. 7922-7928, 2010.
[9]H. He and E. A. Garcia, “Learning from imbalanced data,” IEEE Trans. Knowl. Data Eng., vol. 21, no. 9, pp. 1263–1284, Sep. 2009.
[10]C. Iber, S. Ancoli-Israel, A. Chesson, and S. F. Quan, “The AASM manual for the scoring of sleep and associated events: Rules, terminology and technical specifications,” American Academy of Sleep Medicine, Westchester, 2007.
[11]J. E. Jackson, A User’s Guide to Principal Components, New York: John Wiley and Sons, 1991.
[12]B. Koley and D. Dey “An ensemble system for automatic sleep stage classification using single channel EEG signal,” Computers in Biology and Medicine, vol. 42, no. 12, pp. 1186-1195, 2012.
[13]G. Lu, F. Yang, J. A. Taylor, and J. F. Stein, “A comparison of photoplethysmography and ECG recording to analyze heart rate variability in healthy subjects,” J. of Medical Engineering & Technology, vol. 33, pp. 634–41, 2009.
[14]D. Michie, D. J. Spiegelhalter,, C. C. Tayor, Machine Learning, Neural and Statistical Classification, Ellis Horwood, 1994
[15]A. M. Martinez and A. C. Kak, “PCA versus LDA,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol.23, no. 2, pp. 228-233, 2001.
[16]M. O. Mendez, M. Mattucci, V. Castronovo, L. Ferini-Strambi, S. Cerutti, and A. M. Bianchi, “Sleep staging from heart rate variability: Time varying spectral features and hidden Markov models,” Int. J. Biomed. Eng. Technol., vol. 3, no. 3/4, pp. 246–263, 2010.
[17]K. Pearson, “On lines and planes of closest fit to systems of points in space,” Philosophical Magazine, vol. 2, no. 6, pp. 559-572, 1901.
[18]H. Matsushima, K. Hirose, K. Hattori, H. Sato, and K. Takadama, ”Sleep stage estimation by evolutionary computation using heartbeat data and body-movement,” International Journal of Advancements in Computing Technology, vol. 4, No. 22, pp. 281-290, 2012.
[19]J. P. Niskanen, M. P. Tarvainen, P. O. Ranta-aho, and P. A. Karjalainen, “Software for advanced HRV analysis,” Computer Methods and Programs in Biomedicine, vol. 76, pp. 73-81, 2004.
[20]E. Oropesa, H. L. Cycon and M. Jobert, “Sleep stage classification using wavelet transform and neural network ,” ICSI Technical Report, 1999.
[21]A. Rechtschaffen and A. Kales, Eds., A Manual of Standardized Terminology, Techniques, and Scoring System for Sleep Stages of Human Subjects, U.S. Department of Health, Education and Welfare, 1968.
[22]L. R. Rabiner and B. H. Juang, “An introduction to hidden Markov models,” IEEE ASSP Mag., vol. 3, no. l, pp. 4-16, 1986.
[23]S. J. Redmond and C. Heneghan, “Cardiorespiratory-based sleep staging in subjects with obstructive sleep apnea,” IEEE Trans. Biomed. Eng., vol. 53, no. 3, pp. 485–496, 2006.
[24]S. J. Redmond, P. De Chazal, C. OBrien, S. Ryan, T. W. McNicholas, and C. Heneghan, “Sleep staging using cardiorespiratory signals,” Somnologie, vol. 11, pp. 245–256, 2007.
[25]M. P. Tulppo, T. H. Mäkikallio, T. E. S. Takala, T. Seppänen, and H. V. Huikuri, “Quantitative beat-to-beat analysis of heart rate dynamics during exercise,” American Journal of Physiology-Heart and Circulatory Physiology, vol. 271, pp. 244-252, 1996.
[26]J. Trinder, J. Kleiman, M. Carrington, S. Smith, S. Breen, N. Tan, and Y. Kim, “Autonomic activity during human sleep as a function of time and sleep stage,” J. Sleep Res, vol. 10, no. 4, pp. 253–264, 2001.
[27]G. D. Vito, S. D. R. Galloway, M. A. Nimmo, P. Maas, and J. J. V. McMurray, “Effects of central sympathetic inhibition on heart rate variability during steady-state exercise in healthy humans,” Clinical Physiology and Functional Imaging, vol. 22, pp. 32-38, 2002.
[28]X. Xu and S. Schuckers, “Automatic detection of artifacts in heart period data,” J. Electrocardiol., vol. 34, 2001.
[29]M. Xiao, H. Yan, J. Song, Y. Yang, and X. Yang, “Sleep stages classification based on heart rate variability and random forest,” Biomedical Signal Processing and Control, vol. 8, no. 17, pp. 624–633, 2013.
[30]洪千涵，基於心電訊號之自動情緒辨識演算法之開發，國立成功大學電機工程學系碩士論文，2013。
[31]蘇木春、張孝德，類神經網路、模糊系統以及基因演算法則，第二版，全華，台北，1999。