- 學位論文
應用資料探勘於皮膚疾病之研究與實作
A Study and Implementation of Applying Data Mining on Dermatologic Diseases
摘要
預防勝於治療是預防疾病發生之首要工作，而生病為人類一生中需經歷的人生過程，因此若能藉由預測機制，提前告知使用者共患疾病即將發生的可能性，讓使用者改善可能的致病環境因素，如此就能避免罹患皮膚共患疾病的機率，也免去不必要的發病痛苦與金錢的花費。 本研究旨在建立一個預測系統，於研究中使用中央健康保險局所累積的國人就診紀錄，結合中央氣象局的資料當作探勘資料，並作為實驗資料;本研究針對異位性皮膚炎疾病(Atopic dermatitis)建立一個對應的預測系統，並將關聯法則與TGSP(Time-Gap Sequential Pattern)演算法結合改良出一ARTGSP(Association Rules on Time-Gap Sequential Pattern)演算法，然後將ARTGSP演算法導入本預測系統，並透過Zigbee感測器所測到的使用者環境溫度、濕度，每隔一段時間預測系統自動使用ARTGSP演算法來運算，預測是否會發病與幾天內發病，並警告提醒使用者改善可能的致病因素，讓使用者可避免相關皮膚共患疾病的發生。
並列摘要
Prevention is better than cure is the primary prevention of disease sick need to experience life in the human life process, so if by a prediction mechanism to inform users in advance a total of suffering from the imminent possibility of the disease, allowing users to improve may pathogenic environmental factors, and so be able to avoid the probability of suffering from the skin of comorbidity, but also eliminates the incidence of unnecessary pain and money to spend. This study aims to establish a projection system to study people accumulated by the Central Health Insurance Bureau is used in medical records, combined with the Central Weather Bureau data as exploration of data, experimental data and as; this disease studies for Atopic Dermatitis established a corresponding forecast system, and associated rules and TGSP(Time-Gap Sequential Pattern) algorithm improved ARTGSP(Association Rules on Time-Gap Sequential Pattern) algorithm, and then import the ARTGSP algorithm for forecast system and user environment logging through Zigbee sensor temperature and humidity, every period of time automatically using ARTGSP algorithm for forecast system operation, predict onset and disease within a few days, and warning the user to improve the possible risk factors, Allow users to avoid the occurrence of comorbid diseases of the skin.