請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/65593
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 周俊廷(Chun-Ting Chou) | |
dc.contributor.author | Yi-Yen Wang | en |
dc.contributor.author | 王以彥 | zh_TW |
dc.date.accessioned | 2021-06-16T23:52:39Z | - |
dc.date.available | 2021-02-19 | |
dc.date.copyright | 2020-02-19 | |
dc.date.issued | 2020 | |
dc.date.submitted | 2020-02-17 | |
dc.identifier.citation | [1] Gaurangi Anand Lovekesh Vig Puneet Agarwal Gautam Shroff Pankaj Malhotra,Anusha Ramakrishnan. Lstm-based encoder-decoder for multi-sensor anomalydetection. InICML 2016 Anomaly Detection Workshop.
[2] Syed Mohsen Naqvi Liang Wang Miao Yu, Adel Rhuma and Jonathon Cham-bers. A posture recognition-based fall detection system for monitoring an elderlyperson in a smart home environment.IEEE Transactions on Information Tech-nology in Biomedicine, 16(6), November 2012. [3] Weihua Sheng Chun Zhu and Meiqin Liu. Wearable sensor-based behavioralanomaly detection in smart assisted living systems.IEEE Transactions on Au-tomation Science and Engineering, 12(4), October 2015. [4] Babak Taati Shehroz S. Khan. Detecting unseen falls from wearable devicesusing channel-wise ensemble of autoencoders.Expert Systems with Applications,87(30), November 2017. [5] Mennatallah Amer. Comparison of unsupervised anomaly detection techniques.Master’s thesis, Media Engineering and Technology German University in Cairo,2011. [6] Arindam Banerjee Varun Chandola and Vipin Kumar. Anomaly detection: Asurvey.ACM Computing Surveys, 41(15), July 2009. [7] Raymond T. Ng Markus M. Breunig, Hans-Peter Kriegel and J ̈org Sander. Lof:identifying density-based local outliers. InSIGMOD ’00 Proceedings of the 2000ACM SIGMOD international conference on Management of data. [8] J ̈org Sander Martin Ester, Hans-Peter Kriegel and Xiaowei Xu. A density-basedalgorithm for discovering clusters a density-based algorithm for discovering clus-ters in large spatial databases with noise. InKDD’96 Proceedings of the SecondInternational Conference on Knowledge Discovery and Data Mining. [9] Zhi-Hua Zhou Fei Tony Liu, Kai Ming Ting. Isolation forest. In2008 EighthIEEE International Conference on Data Mining. [10] M. Peng, Y. Li, Z. Zhao, and C. Wang. System architecture and key technologiesfor 5g heterogeneous cloud radio access networks.IEEE Network, 29(2):6–14,March 2015. [11] Yu-Chun Chien Chia-Fu Lee Chih-Wei Ho, Chun-Ting Chou.Unsuper-vised anomaly detection using light switches for smart nursing homes.In2016 IEEE 14th Intl Conf on Dependable, Autonomic and Secure Comput-ing, 14th Intl Conf on Pervasive Intelligence and Computing, 2nd Intl Confon Big Data Intelligence and Computing and Cyber Science and TechnologyCongress(DASC/PiCom/DataCom/CyberSciTech). | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/65593 | - |
dc.description.abstract | 從 1990 年代開始,老年人口佔總人口之比例逐年上升,從 1993 年的 7% 上升至現今的 14% 且持續增加當中,因此,對於年長者之醫療照護將是重要的社會議題。
隨著物聯網(IoT)的發展,許多新的技術和產品被開發出來並應用於改善年長者之居家生活品質,例如使用穿戴式裝置監測血壓及心跳,除了監測之外,一套完整的醫療照護系統也需要能偵測出發生在年長者身上之異常狀況。 但是對於智慧醫療照護系統而言,還有一些重要的問題需解決,首先,系統中佈署之感測器可能會侵犯到使用者之隱私,另外在這些系統中,很能獲得經過標記之資料,導致監督式異常檢測演算法無法使用,而對於非監督式演算法來說,許多超參數需要事先決定,根據超參數設定不同,結果也可能不一樣。 為了解決這些問題,本論文提出了一套基於被動式紅外線感測器之醫療照護系統,來監測年長者之日常生活異常,另外更提出了一套不須事先設定超參數之非監督式異常檢測演算法,此演算法能監測出例如過度異常事件、異常不活躍事件等發生於年長者日常生活中之異常。 | zh_TW |
dc.description.abstract | Since 1990’s, the elderly population has increased significantly. In 1993, the elderly population was 7\%. It has increased to more than 14\% nowadays and is continually increasing. Thus, the healthcare of the elderly is becoming a serious social problem.
With the rising of Internet of Things (IoT), new solutions are being developed to improve the living quality of the elderly at home. For example, blood pressure and heartbeats of the elderly can be monitored with wearable sensors. Besides monitoring, a complete healthcare solution also needs to detect anomalies of the elderly. However, there are some important problems of anomaly detection in smart healthcare systems. First, deployed sensors in these systems may lead to problems such as privacy violation and uncomfortable user experiences. Second, it is difficult to get labeled data in these systems. As a result, supervised anomaly detection algorithms cannot be used. Last, these anomaly detection algorithms have a disadvantage that some hyper-parameters must be decided to detect anomalies. Depends on the choices of these hyper-parameters, results can be different. To solve these problems, a smart healthcare system using passive infrared motion sensors designed to monitor the daily life of the elderly is considered. Moreover, an unsupervised anomaly detection algorithm with no hyper-parameter is proposed in this thesis. This algorithm can detect anomalies such as abnormal active events and inactive events in the daily behavior of the resident. | en |
dc.description.provenance | Made available in DSpace on 2021-06-16T23:52:39Z (GMT). No. of bitstreams: 1 ntu-109-R05942103-1.pdf: 1268521 bytes, checksum: 1082051237d201132590d93bedc8b1e2 (MD5) Previous issue date: 2020 | en |
dc.description.tableofcontents | ABSTRACT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ii
LIST OF TABLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vi LIST OF FIGURES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vii CHAPTER 1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . .1 1.1 Abnormal Conditions for the Elderly . . . . . . . . . . . . . . . . .1 1.1.1 Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 1.1.2 Insomnia . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 1.1.3 Unusual Daily Schedules . . . . . . . . . . . . . . . . . . . .2 1.2 Three Steps in Anomaly Detection. . . . . . . . . . . . . . . . . .2 1.2.1 Step 1: Data Collection . . . . . . . . . . . . . . . . . . . . .2 1.2.2 Step 2: Training/Modeling (Optional) . . . . . . . . . . . . .5 1.2.3 Step 3: Prediction/Detection . . . . . . . . . . . . . . . . . .6 1.2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 1.4 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 1.5 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . . .8 CHAPTER 2 RELATED WORK. . . . . . . . . . . . . . . . . . . . .9 2.1 Supervised Approaches . . . . . . . . . . . . . . . . . . . . . . . . .9 2.1.1 Neural Network-based . . . . . . . . . . . . . . . . . . . . . .9 2.1.2 Computer Vision-based Fall Down Detection . . . . . . . . .10 2.2 Unsupervised Approaches . . . . . . . . . . . . . . . . . . . . . . . .10 2.2.1k-Nearest Neighbors . . . . . . . . . . . . . . . . . . . . . . .11 2.2.2Local Outlier Factor . . . . . . . . . . . . . . . . . . . . . . .12 2.2.3Density Based Spatial Clustering of Applications with Noise13 2.2.4k-Means . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 2.2.5Gaussian Mixture Model . . . . . . . . . . . . . . . . . . . .15 2.2.6Isolation Forest . . . . . . . . . . . . . . . . . . . . . . . . .16 2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 CHAPTER 3 SYSTEM SETTING. . . . . . . . . . . . . . . . . . . .18 3.1 Smart Healthcare System . . . . . . . . . . . . . . . . . . . . . . . .18 3.2 Raw Data Format . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 3.3 Anomalies in Data . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 3.4 Data Transformation Methods . . . . . . . . . . . . . . . . . . . . .22 3.4.1Events in Data . . . . . . . . . . . . . . . . . . . . . . . . . .23 3.4.2Inactive Events . . . . . . . . . . . . . . . . . . . . . . . . .24 3.4.3Active Events . . . . . . . . . . . . . . . . . . . . . . . . . .25 3.4.4Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25 CHAPTER 4 PROPOSED ALGORITHM. . . . . . . . . . . . . . .27 4.1 Hierarchical Clustering . . . . . . . . . . . . . . . . . . . . . . . . .27 4.1.1Distance Measurement . . . . . . . . . . . . . . . . . . . . .27 4.1.2Feature Scaling . . . . . . . . . . . . . . . . . . . . . . . . .29 4.1.3The Problem of Hierarchical Clustering . . . . . . . . . . . .29 4.2 Proposed Method . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 4.2.1Experiment Dataset . . . . . . . . . . . . . . . . . . . . . . .30 4.2.2Finding the Jumping Point . . . . . . . . . . . . . . . . . . .31 4.2.3Remove Unnecessary Merge Steps . . . . . . . . . . . . . . .34 4.2.4Anomalies in Clusters . . . . . . . . . . . . . . . . . . . . . .34 CHAPTER 5 RESULTS AND DISCUSSIONS. . . . . . . . . . . .38 5.1 Results of the Proposed Algorithm . . . . . . . . . . . . . . . . . . .38 5.2 Results of Related Algorithms . . . . . . . . . . . . . . . . . . . . .39 5.2.1Isolation Forest . . . . . . . . . . . . . . . . . . . . . . . . .40 5.2.2K-Nearest Neighbors . . . . . . . . . . . . . . . . . . . . . .40 5.2.3DBSCAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 5.3 Comparisons and Discussions . . . . . . . . . . . . . . . . . . . . . .43 CHAPTER 6 CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . .44 | |
dc.language.iso | zh-TW | |
dc.title | 應用於銀髮居家長照之非監督式異常檢測 | zh_TW |
dc.title | Unsupervised Anomaly Detection for In-Home Elderly Care | en |
dc.type | Thesis | |
dc.date.schoolyear | 108-1 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 逄愛君(Ai-Chun Pang),施吉昇(Chi-Sheng Shih),魏宏宇(Hung-Yu Wei) | |
dc.subject.keyword | 異常檢測,非監督式演算法,物聯網(IoT),智慧醫療照護, | zh_TW |
dc.subject.keyword | Anomaly detection,unsupervised,IoT,smart healthcare, | en |
dc.relation.page | 46 | |
dc.identifier.doi | 10.6342/NTU202000502 | |
dc.rights.note | 有償授權 | |
dc.date.accepted | 2020-02-17 | |
dc.contributor.author-college | 電機資訊學院 | zh_TW |
dc.contributor.author-dept | 電信工程學研究所 | zh_TW |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-109-1.pdf 目前未授權公開取用 | 1.24 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。