- 期刊
- OpenAccess
臺灣地區溺水事件之空間分析－以水域救援統計資料為例
Spatial Analysis of Drowning Incidents in Taiwan-A Case Study of Water Rescue Statistics Data
摘要
目的：本研究旨在探討空間因素對溺水的影響，以提供政府制定防溺政策之參考。方法：研究以2014至2019年消防機關水域救援統計公開資料進行分析，使用Moran's I指標確認空間相依性，並使用地理加權羅吉斯迴歸分析臺灣不同區域對於反應變數是否有不同效果。結果：一、造成溺斃的重要因素包括性別、溺水水域、溺水原因、年齡以及溺水地點與最近醫療單位距離。二、在顯著資料點地區，男性的溺斃率高於女性，淡水水域的溺斃率高於海岸線水域，年齡與溺斃機率呈現正向關係。三、在臺中市及苗栗縣，個人行為因素對於溺斃率的影響較大；而在其他顯著資料點地區，環境危險因素的溺斃率高於個人行為因素。四、在溺水地點與最近醫療單位距離上，除了新竹縣及宜蘭縣外，其他顯著資料點地區，醫療單位距離與溺斃機率成現正向關係。結論：使用地理加權羅吉斯迴歸做為研究模型，對於不同區域的反應變數具有不同效果，能更清楚呈現溺水的空間樣態，可供作為縣市區域擬定防溺措施的參考。創新性：1.以空間分析結果提供政策制定的參考價值。2.考量多個因素的複雜模型。3.為臺灣首篇將地理加權統計應用在溺水事件分析的文章。
並列摘要
Purpose: This study is to investigate the impact of spatial factors on drowning incidents to provide a reference for the government in developing drowning prevention policies. Method: The study analyzed open data of fire department aquatic rescue statistics from 2014 to 2019 and used Moran's I indicator to confirm spatial dependency. A geographically weighted logistic regression (GWLR) was used to analyze whether different regions of Taiwan affect the response variable. Results: 1. the important factors affecting drowning include gender, location, cause of drowning, age, and distance between the drowning location and the nearest medical facility. 2. In significant data point regions, the drowning rate of males is higher than that of females, and the drowning rate in freshwater areas is higher than that of coastal areas. At the same time, age is positively related to drowning probability. 3. In Taichung City and Miaoli County, personal behavioral factors have a greater impact on drowning rates. In contrast, in other significant data point regions, environmental hazard factors have a higher drowning rate than personal behavioral factors. 4. Except for Hsinchu and Yilan counties, distance from medical facilities is positively related to the likelihood of drowning in other significant data point regions. Conclusion: Using geographically weighted logistic regression as a research model has different effects on the response variables in different regions, which can more clearly present the spatial pattern of drowning and provide a reference for counties and cities to formulate drowning prevention measures. Originality/value:1. Provide policy-making with valuable references based on spatial analysis results.2. Consider complex models that take into account multiple factors.3. The first article in Taiwan to apply geographically weighted statistics in drowning analysis.