Developing a Framework of Biomedical Data Classification Methodologies for Telehealthcare Applications

• Main Authors: Hanjun Lin, 林翰俊
• Other Authors: Yeh-Liang Hsu
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/96605767699896855739

博士 === 元智大學 === 機械工程學系 === 101 === This dissertation presents a framework for biomedical data classification methodologies specifically for telehealthcare applications. Biomedical data is categorized into four categories in this study. An example case is demonstrated for each category with the corresponding classification methodology in telehealthcare application. For single-valued vital sign measurements, a rule-based expert system for telehealth application (TES) is demonstrated for the classification of abnormal events, including vital sign abnormality, compliance of measurement and malfunction of hardware devices. For vital sign time series, the AIIA method for classification of heart beats time series into healthy group and disease groups is demonstrated. For human motions and behaviors, a statistical estimation for classification of abnormal and abnormal gaits of Parkinson disease patients is demonstrated. For subjective observation of symptoms, a telehealthcare decision support system (TDSS) for patients recently discharged from hospital, which provides a degree of urgency of returning to the hospital, is demonstrated. This dissertation may not completely cover the classification of all biomedical data. It is intended to provide a structural research framework for biomedical data classification in the telehealth domain. Based on this framework, more example cases can be developed in the field of telehelath.