臺灣博碩士論文加值系統

行為活動識別一直是相當重要的研究議題，從早期利用影像辨識到現今各種不同的自動化識別方法，隨著物聯網與各式雲端應用服務的興起，結合智慧行動裝置帶動的電子科技微型化技術，使得穿戴式裝置在行為活動識別領域之應用日趨重要。根據市調機構IDC的統計，穿戴式裝置之普及率將以每年23%的幅度成長，在如此廣泛的使用程度下，如何透過穿戴式科技進行行為活動偵測，以便更加快速的協助人類在特定環境下完成指定任務是行為活動識別最重要的目標。針對行為活動識別方面的研究。截至目前，已經有許多學者開始以智慧型手機當作感測來源對行為活動進行預測。相較於智慧型手機而言，穿戴式裝置之人體分佈範圍廣；也因為功能較少，電池壽命長；加上老年族群對智慧型手機之依賴程度普遍較低，因此利用穿戴式裝置作為行為活動識別來源會更加具備效益。
本研究以OPPORTUNITY行為活動識別資料集為原始資料，首先利用資訊獲利、資訊增益比與oneR演算法，篩選各行為活動之重要人體部位，透過三軸加速度感測器，與行為活動相互驗證。再以四種不同之決策樹演算法：ID3、C4.5、分類回歸樹、卡方自動交互檢視法與類神經網路建構預測模型及準確率比較，並利用關聯規則演算法-Apriori找出與行為活動相關之人體部位組合，進一步探究各部位之關聯性。本研究顯示右側膝蓋為判別行為活動之最重要部位，而針對平躺行為活動而言，其識別準確率為所有行為活動中最高，期許能作為未來穿戴式裝置在行為活動識別研究之相關依據，並與不同領域相結合，協助人類完成不同工作。

Human Activity Recognition has been a very important issue since image recognition to automatic identification, with the IOT and Cloud computing services accompany with technology miniaturization, wearable devices has a rapid importance growth in activity recognition. According to IDC statistics, wearable devices will have an annual growth of twenty three percent in the last five years. With such a wide degree of usage, the way using these wearable devices to assist human in their task more efficiently and convenient is the most important goal of Human Activity Recognition. Most studies of Human Activity Recognition uses smartphones as data source. Compared to smartphones, wearable devices has a widely distribution on human body. With less function uses, long battery life, and the reliance of smartphones on elderly people, there will be more benefits using wearable devices’ sensors on behavior identification.
The OPPORTUNITY Activity Recognition Dataset is used in this study to propose a method of data mining which can identify different locomotion. This study is divided into three parts. The first part is to use different feature selection method to select the important part of human body which have the most relation to locomotion. The Information Gain, Gain Ratio and OneR algorithms are used in this part, the result can be verified after the selection. In the second part, ID3, C4.5, CART, CHAID and Artificial Neural Network are used to construct an intelligent classification model with the comparison of accuracy between these models. After the two parts, Apriori is used to find potential association rule combined by different parts of the human body. This study shows that the most important part of human body to identify activity is the alccelerometer worn on the right knee, also having powerful associations in most locomotion. Among all the locomotion, models to identify locomotion of lie has the highest accuracy which is better performed in C4.5 and Artificial Neural network. Expecting the result can be a basis in Human Activity Recognition research area, also making a progress in proactively assisting human on their task.

誌謝 i
摘要 ii
ABSTRACT iii
目錄 iv
圖目錄 vii
表目錄 x
第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 3
1.3 研究架構 3
1.4 研究限制 5
第二章 文獻探討 6
2.1 三軸加速度感測器(Triaxial Accelerometer) 6
2.2 特徵選取(Feature Selection) 8
2.3 機器學習(Machine Learning) 9
2.4 關聯規則(Association Rule) 11
第三章 研究架構與方法 14
3.1 資料來源(Data Source) 15
3.2 研究環境(Research Environment) 21
3.2.1 WEKA資料探勘軟體 21
3.2.2 Statistica統計軟體 22
3.2.3 R統計分析軟體 23
3.3 人體部位篩選(Feature Selection) 24
3.3.1 資訊獲利(Information Gain) 25
3.3.2 資訊獲利比例(Gain Ratio) 28
3.3.3 OneR演算法(OneR Algorithm) 29
3.4 決策樹演算法(Decision Tree) 31
3.4.1 ID3演算法(ID3 Algorithm) 31
3.4.2 C4.5演算法(C4.5 Algorithm) 33
3.4.3 CART演算法(CART Algorithm) 35
3.4.4 CHAID演算法(CHAID Algorithm) 35
3.4.5 決策樹演算法之比較 37
3.5 類神經網路(Artificial Neural Network) 37
3.6 Apriori演算法(Apriori Algorithm) 44
第四章 研究結果 46
4.1 人體部位篩選與分類模型驗證 46
4.2 關聯規則分析結果 74
第五章 結論與未來展望 90
參考文獻 91

參考文獻
[1]LukasHorak, "植基於行動裝置之行為辨識及能量消耗監測整合系統," 國立成功大學醫學資訊研究所碩士論文, 2014.
[2]丁一賢 與 陳牧言, 資料探勘: 滄海書局, 2006.
[3]王鈺姍, "以方法目的鏈探討穿戴式裝置使用者之價值結構," 國立台灣科技大學商業管理學系碩士論文, 2016.
[4]宋佩栩, "一個使用環場攝影機並結合個人資訊的客製化跌倒偵測系統," 中原大學電子工程研究所碩士論文, 2006.
[5]林振禾, "在智慧空間下以貝氏網路為基礎的行為活動偵測分析," 國立中正大學資訊工程所碩士論文, 2009.
[6]洪甘錠, "RFID科技整合影像分析於老人跌倒之研究," 輔仁大學資訊管理學系碩士論文, 2007.
[7]袁梅宇, 王者歸來:WEKA機器學習與大數據聖經, 2015.
[8]彭淯湘, "基於關聯式分類法在手機上進行人類行為活動辨識," 國立交通大學資訊科學與工程研究所碩士論文, 2014.
[9]黃閔珊, "於行動裝置上具持續性和個人化之日常行為辨識系統設計," 國立成功大學資訊工程學系碩士論文, 2014.
[10]賴建翔, "運用三維特徵於智慧型手機上進行人類行為活動辨識," 國立交通大學網路工程研究所碩士論文, 2015.
[11]鍾宜岑, "應用於動態背景中的移動物體影像之偵測與即時追蹤系統," 國立交通大學電機學院IC設計產業專班碩士論文, 2007.
[12]簡禎富 與 許嘉裕, 資料挖礦與大數據分析: 前程文化, 2014.
[13]J. M. Alexandre Santosa, António Costaa, M. João Nicolau, "Internet of Things and Smart Objects for M-health Monitoring and Control," Procedia Technology, vol. 16, pp. 1351-1360, 2014.
[14]L. BREIMAN, "Bagging Predictors," Machine Learning, vol. 24, pp. 123-140, 1996.
[15]T. Chen, Zhang, X., Jin, S., & Kim, O., "Efficient classification using parallel and scalable compressed model and its application on intrusion detection," Expert Systems with Applications, vol. 41, pp. 5972-5983, 2014.
[16]A. H. T. Cheol-Hong Min, "Automatic characterization and detection of behavioral patterns using linear predictive coding of accelerometer sensor data," presented at the 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, Buenos Aires, 2010.
[17]A. C. Daniel Roggen, Mirco Rossi, Thomas Holleczek, Gerhard Tröster, Paul Lukowicz, Gerald Pirkl, David Bannach, Alois Ferscha, Jakob Doppler, Clemens Holzmann, Marc Kurz, Gerald Holl, Ricardo Chavarriaga, Hesam Sagha, Hamidreza Bayati, and José del R. Millàn, "Collecting complex activity data sets in highly rich networked sensor environments," presented at the Seventh International Conference on Networked Sensing Systems Kassel, Germany, 2010.
[18]A. N. Fleury, N; Vacher, M; Glasson, H. Seri, J.-F., "Sound and speech detection and classification in a Health Smart Home," Engineering in Medicine and Biology Society, 2008.
[19]J. V. G. Bieber, B. Urban, " Activity recognition for everyday life on mobile phones," presented at the Proceedings of the 5th International on Conference Universal Access in Human-Computer Interaction, Part II: Intelligent and Ubiquitous Interaction Environments, 2009.
[20]D. Hebb, The Organization of Behavior: A Neuropsychological Theory, 1949.
[21]J. J. Hopfield, "Neural networks and physical systems with emergent collective computational abilities," Biophysics, vol. 79, pp. 2554-2558, 1982.
[22]S. M. J. Frank, D. Precu, "Activity recognition with mobile phones," presented at the Proceedings of European Conference on Machine Learning 2011 Demo Session, 2011.
[23]T. p. a. d. J. Yang, "motion recognition using simple acceleration features with mobile phones," presented at the Proceedings of the 1st International Workshop on Interactive Multimedia for Consumer Electronics, 2009.
[24]G. V. Kass, "An Exploratory Technique for Investigating Large Quantities of Categorical Data," Journal of the Royal Statistical Society, vol. 29, pp. 119-127, 1980.
[25]D. Z. L. Sun, B. Li, B. Guo, S. Li, "Activity recognition on an accelerometer embedded mobile phone with varying positions and orientations," presented at the Proceedings of the 7th International Conference on Ubiquitous Intelligence and Computing, 2010.
[26]J.-F. S. a. S. C. Michel Vacher, "Sound Classification in a smart room environment: an approach using GMM and HMM Methods," presented at the The 4th Conference on Speech Technology and Human-Computer Dialogue, Romania, 2007.
[27]M. A. L. Oscar D. Lara, "A Survey on Human Activity Recognition using Wearable Sensors," Communications Surveys & Tutorials, vol. 15, pp. 1192-1209, 2013.
[28]J. R. Quinlan, "Induction of decision trees," Machine Learning, vol. 1, pp. 81-106, 03/01 1986.
[29]J. R. Quinlan, "C4.5: programs for machine learning," Morgan Kaufmann Publishers Inc, 1993.
[30]J. R. Quinlan, "Improved Use of Continuous Attributes in C4.5," Journal of Artificial Intelligence Research, vol. 4, pp. 77-90, 1996.
[31]H. M. Rakesh Agrawal, Ramakrishnan Srikant, Hannu Toivonen, A. Inkeri Verkamo, "Fast discovery of association rules," in Advances in knowledge discovery and data mining, ed, 1996, pp. 307-328.
[32]R. S. Rakesh Agrawal, "Fast Algorithms for Mining Association Rules in Large Databases," presented at the Proceeding VLDB 94 Proceedings of the 20th International Conference on Very Large Data Bases, 1994.
[33]R. A. Ramakrishnan Srikant, "Mining Generalized Association Rules," presented at the VLDB 95 Proceedings of the 21th International Conference on Very Large Data Bases, 1995.
[34]H. S. Ricardo Chavarriaga, Alberto Calatroni, Sundaratejaswi Digumarti, Gerhard Tröster, José del R. Millán, Daniel Roggen, "The Opportunity challenge: A benchmark database for on-body sensor-based activity recognition," Pattern Recognition Letters, 2013.
[35]F. Rosenblatt, "The perceptron: A probabilistic model for information storage and organization in the brain," Psychological Review, vol. 65, pp. 386-408, 1958.
[36]J. G. C. Ryszard S. Michalski, Tom M. Mitchell, Machine Learning: An Artificial Intelligence Approach, 1985.
[37]D. W. M. S. Consolvo, T. Toscos, M.Y. Chen, J. Froehlich, B. Harrison, P. Klasnja, A. LaMarca, L. LeGrand, R. Libby, I. Smith, J.A. Landay, "Activity sensing in the wild: a field trial of ubifit garden," presented at the Proceedings of the Twenty-sixth Annual SIGCHI Conference on Human Factors in Computing Systems, 2008.
[38]C. E. Shannon, "A mathematical theory of communication," ACM SIGMOBILE Mobile Computing and Communications Review, vol. 5, pp. 3-55, 2001.
[39]M. E. S. Stephen J. Redmond, Michael R. Narayanan, Stephen R. Lord, Sergio Cerutti and Nigel H.Lovell, "Automatic Segmentation of Triaxial Accelerometry Signals for Falls Risk Estimation," presented at the 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, Buenos Aires, 2010.
[40]J. G. T. Brezmes, J. Cotrina, "Activity recognition from accelerometer data on a mobile phone," presented at the Proceedings of the 10th International Work-Conference on Artificial Neural Networks: Part II: Distributed Computing, Artificial Intelligence, Bioinformatics, Soft Computing, and Ambient Assisted Living, 2009.
[41]S. C. T. Choudhury, B. Harrison, J. Hightower, A. LaMarca, L. LeGrand, A. Rahimi, A. Rea, G. Borriello, B. Hemingway, P.Pedja Klasnja, K. Koscher, J.A. Landay, J. Lester, D. Wyatt, D. Haehnel, "The mobile sensing platform: an embedded activity recognition system," Pervasive Computing, vol. 7, pp. 32-41, 2008.
[42]A. C. Thomas Petzsche, Marine Dumont, "Triaxial Accelerometer, High Frequency Measurement and Temperature Stability Considerations," in Special Topics in Structural Dynamics. vol. 6, ed, 2014, pp. 333-340.
[43]W. P. Warren S. McCulloch, "A logical calculus of the ideas immanent in nervous activity," The bulletin of mathematical biophysics, vol. 5, pp. 115-133, 1943.
[44]I.-J. K. Y.-J. Hong, S.C. Ahn, H.-G. Kim, "Mobile health monitoring system based on activity recognition using accelerometer," Simulation Modelling Practice and Theory, vol. 18, 2010.
[45]J. L. Y. Wang, M. Annavaram, Q.A. Jacobson, J. Hong, B. Krishnamachari, and N. Sadeh, "A framework of energy efficient mobile sensing for automatic user state recognition," presented at the Proceedings of the 7th International Conference on Mobile Systems, Applications, and Services, 2009.
[46]J. Yang, & Honavar, V., "Feature subset selection using a genetic algorithm," Feature extraction, construction and selection pp. 117-136, 1998.
[47]D. B. Yucheng Low, Joseph Gonzalez, Carlos Guestrin, Aapo Kyrola, Joseph M. Hellerstein, "Distributed GraphLab: a framework for machine learning and data mining in the cloud," Proceedings of the VLDB Endowment, vol. 5, pp. 716-727, 2012.
[48]Zhang, X., Hu, B., Ma, X., Moore, P., & Chen, J. (2014). Ontology driven decision support for the diagnosis of mild cognitive impairment. Computer methods and programs in biomedicine, 113(3), 781-791.
[49]Zhou, Q., Wu, Y., Chan, C. W., & Tontiwachwuthikul, P. (2011). Modeling of the carbon dioxide capture process system using machine intelligence approaches. Engineering Applications of Artificial Intelligence, 24(4), 673-685.