隨著科技不斷演進，醫療技術越來越成熟，人口老化現象逐年上升。但因醫療照護費用昂貴及個人隱私等問題，現在的老年人較希望可在自家自行休養。而現今在人們生活中運用智慧感測技術越來越能夠被人們接受，為了能滿足老年人所希望的生活方式，可透過在居家環境及家電上所安裝的感測器即時感測老年人狀況，還可透過機器學習及分析技術了解老年人平常的生活模式。此類結果未來除了能夠運用於智慧家庭來輔助觀察及照顧老年人，同時還可了解老年人在家中是否有異常狀況發生之情形。
由於過往行為模式探勘大部份在進行關聯分析時都以固定支持度值在進行關聯性篩選，這意味著老年人在智慧家庭中每天都是以相同模式在進行生活。但在現實中每天一成不變的生活模式不太現實，生活難免會因不同事物有些微的改變。因此，本篇論文提出了在智慧家庭中透過自適性行為模式探勘進行週期性活動辨識，藉此達到老年人在家生活習性之判別。為了達到此目的，本研究會先將感測器資料轉換為事件序列，此序列在依時間區間劃分出不同情節，並以先驗演算法執行頻繁情節探勘(Frequent Episode Mining)。在此情節探勘中所應用之支持度會以自適性的方式進行，確保情節探勘可依每天些微生活習性的改變而自我調整。此自適性支持度將利用歷史資料結合加權移動平均，並續運用最大似然法與高斯分布找出具有週期性的行為。透過實驗證明自適性支持度有助於行為探勘依據老年人生活習性上的變化進行調整，而本論文與固定支持度相比平均提高行為辨識準確度7.33%。

With the continuous improvement of science and technology, medical technology is becoming more and more mature, and the phenomenon of increasing population is increasing year by year. However, due to the high cost of medical care and personal privacy, many older people wish to recuperate at home. Nowadays, the technology of using intelligent sensing in people's lives is more and more accepted. To meet the lifestyle of the elderly, the sensors installed in the home environment and home appliances can not only detect the elderly Real-time conditions can even be machine-learned and the resulting technology can be applied to smart homes in the future to assist elderly observation and care, and at the same time understand whether the elderly have abnormal behaviors at home.
Because most of the past behavior pattern explorations used associative analysis to perform correlation screening with a fixed support value, this means that the elderly live in the same mode every day in smart families. However, in reality, the unchanging daily life pattern is not realistic, and life will inevitably change slightly due to different things. Therefore, this paper proposes the identification of periodic living activities in smart homes through the exploration of adaptive behavior patterns to achieve the elderly's home habits. First, the sensor data is converted into a sequence of events. This sequence is divided into different episode according to time intervals, and frequent episode mining are performed using apriori algorithm. The support applied in this episode mining will be carried out adaptively, ensuring that the plot exploration can be Self-adjustment according to changes in daily life habits. To achieve this, this adaptive support will use historical data combined with weighted moving averages and continue to use the maximum likelihood estimation and Gaussian distribution to find periodic behavior. Through experiments, it is proved that adaptive support is helpful for behavioral exploration to adjust according to the changes in the elderly's living habits. Compared with fixed support, this paper improves the average accuracy of activity recognition by 7.33%.

目錄
第一章 、緒論 1
1.1 前言. 1
1.1.1 物聯網歷史與發展趨勢 1
1.1.2 智慧家庭應用發展 2
1.1.3 面對人口高齡化議題與現象 6
1.2 研究動機與目的 7
1.2.1 老年人智慧居家照護議題 7
1.2.2 行為辨識技術應用 8
1.2.3 研究目的 8
1.3 論文架構 9
第二章 、文獻探討與技術分析 10
2.1 行為辨識 10
2.2 常見應用於行為辨識技術 12
2.2.1 隱馬可夫模型 13
2.2.2 單純貝氏分類器 14
2.2.3 支援向量機 15
2.2.4 決策樹 16
2.2.5 先驗演算法 17
2.2.6 FP-growth 18
2.2.7 可擴展的混和行為辨識系統 20
2.3 小結. 20
第三章 、智慧家庭環境中透過自適性行為模式探勘進行週期性活動辨識 21
3.1 系統架構 22
3.2 蒐集感測器資料 23
3.3 資料前處理 25
3.4 行為探勘 26
3.4.1 事件序列劃分 31
3.4.2 自適性支持度 33
3.4.3 頻繁情節探勘 36
3.4.4 週期性分析 38
第四章 、實驗測試與分析 41
4.1 實驗目的 41
4.2 資料集介紹 41
4.3 實驗參數 47
4.4 實驗結果 48
4.5 實驗結果比較 55
第五章 、結論與未來展望 58
5.1 研究結論 58
5.2 未來展望 59
第六章 、參考文獻 60

[1]每日頭條，[Online]. Available：https://kknews.cc/zh-tw/tech/nxgo9e8.html
[2]科學月刊， [Online]. Available：http://scimonth.blogspot.tw/2014/05/blog-post_3117.html
[3]每日頭條，[Online]. Available：https://kknews.cc/zh-tw/tech/yzb86mg.html
[4]瘋科技， [Online]. Available：https://www.cool3c.com/article/144290
[5]3S Market「全球智慧科技應用」市場資訊網， [Online]. Available：https://3smarket-info.blogspot.com/2019/01/2019_8.html
[6]集微網，紫光:2020年手機晶片份額世界第一；2020物聯網產值破兆美元，[Online]. Available：https://wechatinchina.com/thread-25819-1-1.html
[7]MoneyDJ，IDC：智慧居家裝置出貨爆發，未來 5 年維持兩位數成長，[Online]. Available：https://ccc.technews.tw/2018/10/02/idc-smart-home-shipments/
[8]pixabay，[Online]. Available：https://pixabay.com/images/search/smart%20home/
[9]內政部統計年報，[Online]. Available：https://www.moi.gov.tw/files/site_stuff/321/2/year/year.html
[10]World Population Prospects -The 2017 Revision- Volume II: Demographic Profiles
[11]World Population Prospects - The 2015 Revision - Volume II : Demographic Profiles
[12]高齡社會已到！臺灣老年人口比率達 14.05% 朝超高齡社會前進中，老伴ㄦ，[Online]. Available：https://www.crew.com.tw/2018/taiwan-turned-aged-society-march-2018/
[13]M. Marmot, "Fair Society, Healthy Lives, in Inequalities in Health:Concepts, Measures, and Ethics, " New York: Oxford University Press,pp. 282, 2013.
[14]K. Kinsella, J. Beard, and R. Suzman, "Can Populations Age Better,Not Just Live Longer?, " Generations 37.1, pp. 19-26, 2013.
[15]AARP, Available: https://www.aarp.org/
[16]L.Y. Beng, S. Ramadass, S. Manickam and T. S. Fun,” A Survey of Intrusion Alert Correlation and Its Design Considerations,” IETE Technical Review 31:3, pages 233-240, 2014.
[17]每日頭條，機器學習之隱馬爾可夫模型 (HMM)，[Online]. Available：https://kknews.cc/zh-tw/science/p44nja2.html
[18]L. Rabiner ,B. Juang “An introduction to hidden Markov models, ” IEEE ASSP Magazine, vol 3, no 1, pp.4-16, 1986.
[19]L.R. Rabiner ,“A tutorial on hidden Markov models and selected applications in speech recognition, ” Proceedings of the IEEE, vol 77, no. 2, pp.257-286, 1898.
[20]S. Shaily ,V. Mangat ,“The Hidden Markov Model and its application to Human Activity Recognition, ” International Conference on Recent Advances in Engineering & Computational Sciences (RAECS), 2015.
[21]C. Wang , Y. Xu , J. Zhang and W. Yu” SW-HMM: A Method for Evaluating Confidence of Smartphone-Based Activity Recognition,” IEEE Trustcom/BigDataSE/ISPA, pp.2086-2091, 2016.
[22]T.V. Duong , H.H. Bui , D.Q. Phung , S. Venkatesh ”Activity recognition and abnormality detection with the switching hidden semi-Markov model, ” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005.
[23]維基百科，單純貝式分類器，[Online]. Available：https://zh.wikipedia.org/wiki/%E6%9C%B4%E7%B4%A0%E8%B4%9D%E5%8F%B6%E6%96%AF%E5%88%86%E7%B1%BB%E5%99%A8
[24]X. Yang , A. Dinh and L. Chen,” A wearable real-time fall detector based on Naive Bayes classifier,” IEEE CCECE ,2010.
[25]支持向量機（Support Vector Machine, SVM），[Online]. Available：https://medium.com/@aitmr1234567890/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%A9%9F-support-vector-machine-svm-b3691f82b96
[26]Fleury, N. Noury and M. Vacher,” Supervised classification of activities of daily living in health smart homes using SVM,” Annual International Conference of the IEEE Engineering in Medicine and Biology Society ,pp. 6099-6102,2009.
[27]維基百科，單純貝式分類器，[Online]. Available：https://zh.wikipedia.org/wiki/%E5%86%B3%E7%AD%96%E6%A0%91
[28]L. Fan, Z. Wang, H. Wang, “Human Activity Recognition Model Based on Decision Tree, ” International Conference on Advanced Cloud and Big Data, 2013.
[29]R. Agrawal and R. Srikant, “ Fast Algorithms for Mining Association Rules ,” 1994.
[30]C. Yadav, S. Wang and M. Kumar,” An approach to improve apriori algorithm based on association rule mining,” Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT), pp.1-7, 2013.
[31]J. Wen, M. Zhong and Z. Wang,” Activity recognition with weighted frequent patterns mining in smart environments,” Expert Systems with Applications, vol. 42, pp. 6423-6432, 2015.
[32]科技大數據，機器學習實戰項目- FP-growth算法來高效發現頻繁項集，每日頭條，[Online]. Available：https://kknews.cc/code/8pxzm9q.html
[33]Amine Lotfi Bourbia ; Heesuk Son ; Byoungheon Shin ; Taehun Kim ; Dongman Lee ; Soon J. Hyun,” Temporal Dependency Rule Learning Based Group Activity Recognition in Smart Spaces,” IEEE 40th Annual Computer Software and Applications Conference, pp. 658-663, 2016.
[34]T. L. M. van Kasteren, G. Englebienne, and B. J. A. Kröse, “Human Activity Recognition from Wireless Sensor Network Data: Benchmark and Software,” In the book Activity Recognition in Pervasive Intelligent Environments of the Atlantis Ambient and Pervasive Intelligence series, Atlantis Press, pp. 165-186 , 2010.
[35]E. O. Heierman, M. Youngblood, and D. J. Cook, “Mining temporal sequences to discover interesting patterns,” KDD Workshop on miningtemporal and sequential data, 2004.
[36]J. Soulas, and P. Lenca, “Periodic episode discovery over event streams,” Portuguese Conference on Artificial Intelligence, pp. 547-559, 2015.
[37]TCP協定，博碩文化，[Online]. Available：http://dns2.asia.edu.tw/~wzyang/slides/info_net/info_B/CH10TCP.pdf
[38]Max行銷誌，[Online]. Available：https://www.maxlist.xyz/2018/11/03/python_apriori/
[39]Z. Li, J. Han, B. Ding, and R. Kays, “Mining periodic behaviors of object movements for animal and biological sustainability studies,” Data Mining and Knowledge Discovery, pp. 355-386, 2012.
[40]B. Valérie and J. Bouisson, "Vieillesse et routinisation: une revue de la question." Psychologie & NeuroPsychiatrie du vieillissement, vol 6, pp.235-243, 2008.
[41]陳鍾誠 (2009年12月05日)，最大似然法則 (Maximum Likelihood)，陳鍾誠的網站，[Online]. Available：http://ccckmit.wikidot.com/st:maximumlikelihood
[42]J. Soulas, P. Lenca and A. Thépaut, ” Unsupervised discovery of activities of daily living characterized by their periodicity and variability,” Engineering Applications of Artificial Intelligence, vol 45, pp. 90-102, 2015.
[43]G. Azkune and A. Almeida,” A Scalable Hybrid Activity Recognition Approach for Intelligent Environments,” IEEE Access, vol 6, pp. 41745-41759, 2018.
[44]P. Rashidi, D. J. Cook, "COM: A method for mining and monitoring human activity patterns in home-based health monitoring systems", ACM Trans. Intell. Syst. Technol., vol. 4, no. 4, pp. 64-1-64-20, 2013.