| DC 欄位 |
值 |
語言 |
| DC.contributor | 應用地質研究所 | zh_TW |
| DC.creator | 孫綺謙 | zh_TW |
| DC.creator | Chi-Chien Sun | en_US |
| dc.date.accessioned | 2020-8-18T07:39:07Z | |
| dc.date.available | 2020-8-18T07:39:07Z | |
| dc.date.issued | 2020 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=107624008 | |
| dc.contributor.department | 應用地質研究所 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 位在台灣東北部的蘭陽平原地下水儲量豐富，並廣泛的使用在農業灌溉、漁業養殖、家庭用水甚至是飲用水中。然而，部分地下水中的砷濃度超過世界衛生組織的飲用水水質標準且呈現空間變異性，飲用砷污染水造成的健康風險也因地區而不同；因此評估地下水砷污染的空間分佈在水資源保護及公共衛生管理上是很重要的。研究的目的是以倒傳遞訓練法的類神經網路（BPNN）分析在蘭陽平原地下水砷污染之空間變異性，並將結果與普通克利金法(OK)比較，透過交叉驗證的方式將資料分成三組來呈現結果，BPNN及OK的平均決定係數分別為R^2=0.55及R^2=0.49，而平均的均方根誤差為RMSE=0.49及RMSE=0.54。由於BPNN比OK更能準確地預測地下水砷濃度，將其作為一種可靠的地下水污染空間繪圖工具，並根據美國環境保護署提出的致癌風險(Target Risk)及非致癌風險(Hazard Quotient)轉換為健康風險分布圖，用以劃定地下水砷污染的高風險地區，優先考慮對不安全的地下水進行更深入監控；此外，政府需要開發安全的水資源，以替代具有風險的地下水。 | zh_TW |
| dc.description.abstract | Groundwater reserves in the Lanyang plain are abundant and widely used for agricultural irrigation, aquaculture farming, domestic and drinking. However, some groundwater arsenic (As) concentration far exceeds the drinking water quality standard of the World Health Organization (WHO) and exists a spatial variability. For protecting the water resource and improving the public health management, it is necessary to consider the spatial variability of groundwater As contamination. The purpose of this study is to apply back-propogation neuron network (BPNN) methods to carry out spatial mapping of the groundwater As concentration and compare with geostatistical Ordinary Kriging (OK) method. Cross validation is used to distribute the monitoring data into three sets in order to reveal the predicting performance. The results show that the average determination coefficients (R^2) of cross validation for As concentrations obtained with BPNN and OK are 0.55 and 0.49, respectivly, and the average root mean square error (RMSE) are 0.49 and 0.54. Considering that the BPNN can yield a higher correct than OK, it is recommended as a reliable method for spatial mapping the groundwater contamination. Therefore, the As concentrations estimated by BPNN are transformed to the associated human health risk based on the hazard quotient (HQ) and target risk (TR) established by the U.S. Environmental Protection Agency. The spatial maps of the groundwater contamination can be used to demarcated to describe the areas that residents are at high risk due to the ingestion of As containing groundwater, prioritize the areas where more intensive monitoring of unsafe groundwater quality is required. Moreover, the government needs to develop safe water resources as alternatives to using unsafe groundwater. | en_US |
| DC.subject | 倒傳遞類神經網路 | zh_TW |
| DC.subject | 地下水砷污染 | zh_TW |
| DC.subject | 健康風險 | zh_TW |
| DC.subject | 蘭陽平原 | zh_TW |
| DC.subject | back-propogation neuron network | en_US |
| DC.subject | groundwater arsenic contamination | en_US |
| DC.subject | health risk | en_US |
| DC.subject | Lanyang Plain | en_US |
| DC.title | 以人工智慧為基礎進行地下水情勢的空間地圖繪製與公共衛生管理實務應用：蘭陽平原飲用含砷地下水的健康風險空間變異分析 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | An artificial intelligence approach for spatial mapping of groundwater occurrence and its application of public health management: Spatial health risk assessment associated with ingestion of arsenic-affected groundwater in the Lanyang Plain | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |