- 學位論文
利用K均值法與自組織映射圖分析心律指數支援遠距健康照護
Using K-means and Self-Organizing Map to Analyze Heart Rate for Supporting Tele-Healthcare
摘要
近年來老年人口的快速成長，以至於老人照護的相關議題也跟著被重視了起來，隨著科技與資訊技術的日新月異，使得我們可以在不久的將來透過網際網路做即時的遠距健康照護。有別於傳統的醫療照護儀器，遠距健康照護為老人在自己的家裡透過網際網路與照護系統進行長期、非察覺性的健康監測。但在遠距監測時，受測者的生理指數範圍將會比目前所定義的健康正常指數範圍來得大，並且具有許多因日常活動所造成的起伏變化，所以本研究提出一個有學習能力的生理分析系統。此系統在學習階段，將網際網路所傳送過來的老人健康生理指數利用叢集分析中的K均值法與類神經網路中的自組織映射圖(Self-Organizing Map, SOM)，進行學習與分析，再由醫師或照護人員依照學習後的分類結果詢問受測者每組類別中所做過的日常活動。這些個人專屬的生理指數資料，可用來解決因為照護階段所測得指數範圍擴大過度提出警告的問題；並且當受測者健康狀況有惡化徵兆時即能做出正確的預測，告知醫護人員以便做及時的處理，以達到老人健康照護的目的。本論文的重點在生理指數的學習與分析，實驗結果說明所提方法之可行性。
並列摘要
The population of elders has been growing fast in recent years. Therefore, relevant issues of the elders’ healthcare are aroused and stressed. With the advances in computer and network technologies, the real-time tele-healthcare through internet will be possible in the near future. By transmitting vital signs through internet, the elders can be cared at home. It carries out long-term and unaware health monitoring. For long-term health monitoring, the health ranges of the vital signs may be larger than the predefined normal ranges due to activities. For this reason, this research proposes a tele-healthcare system with learning abilities to learn the class diagram of vital signs for each elder by the K-means algorithm and self-organizing map. Then the doctors or healthcarers ask the elder what activities were done for each class of vital signs. The purpose is to estimate the daily normal health range of the elder. After the learning process, when the healthcare system receives the elder’s signals beyond the learned normal range, it calls the healthcarers to care the elder. Hence distant healthcare can be achieved. Experimental results show the feasibility of the proposed method.