空氣汙染物與疾病關聯性之研究與利用深度學習預測疾病;A Study of Correlation between Air Pollutants and Diseases and Diseases Prediction by Deep Learning

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/77602

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/77602

Title:	空氣汙染物與疾病關聯性之研究與利用深度學習預測疾病;A Study of Correlation between Air Pollutants and Diseases and Diseases Prediction by Deep Learning
Authors:	陳俞毓;Chen, Yu-Yu
Contributors:	資訊工程學系
Keywords:	健保資料庫;空氣汙染;分析系統;疾病預測;NHIRD;Air pollution;analysis system;disease prediction
Date:	2018-07-23
Issue Date:	2018-08-31 14:49:39 (UTC+8)
Publisher:	國立中央大學
Abstract:	近年來，環境因素對人類的影響不斷增加，這現象儼然成為一個棘手的問題。在此情形下，很多研究者投入此方向並表明了許多不同種類的疾病與環境因素之間的聯繫，而在這些環境因素中空氣污染和水質之討論尤其眾多。與此同時，環境因素對個體的影響也不盡相同，在這種情形下個體所產生的疾病也不盡相同。因此，在本研究中，我們建立一個自動化分析系統來分析任何疾病與環境因素之間的關係，並構建一個基於深度學習併考量空氣狀態或水質狀態的疾病預測模型。因為在水質部分並無即時值可供我們建構平台，所以平台主要關注在空氣汙染物的部分，而結果表明，儘管我們的疾病與空氣汙染物關聯之自動化系統的分析結果與以往的研究結果有一定的差異，但它們之間的相似的部分仍然佔了大多數。在疾病預測方面，我們在總體的疾病預測上有較好的表現。;In recent year, the influence on human beings of environmental factors increasing has been a hot potato and there was a lot of research has shown the association between different kinds of diseases and environmental factors, especially air pollutants and water quality. Meanwhile, influence on individuals by environmental factors are not all the same, and the diseases they get are different. Therefore, in this study, we want to implement an automatic system to analyze the relationship between any disease on Longitudinal Health Insurance Database (LHID) and environmental factors and construct deep learning-based models for diseases prediction incorporating air status or water quality status. However, there is no instantaneous value of water quality, so we focus on air pollutants on automatic analysis system. The results show that even though there are some differences between the analytical results from our system and the previous research, the similarities between them are in majority. In diseases prediction, we show high performance on the overall forecast. Our models considered medical information from LHID, incorporating air pollutants, location information, and water quality. The accuracy among these four features is 89.49%, 89.59%, 89.59%, and 89.56% separately. In short, incorporating these environmental factors can improve the accuracy of deep learning-based diseases prediction actually.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	144	View/Open

社群 sharing

Loading...