基于聚类分析的集中式供水水质安全风险评价

发布时间：2018-05-24 18:23

本文选题：风险评价 + 集中式供水　；参考：《环境与健康杂志》2017年05期

【摘要】：目的探讨适宜于集中式供水领域特点的风险评价方法。方法选择我国1 302座集中式供水设施,并计算其风险指数。采用两阶段聚类计算最优分类数。采用系统聚类和K-means聚类两种方法,并以最优分类数和标准推荐的三分类分别对研究对象的风险指数进行聚类分析。结果两阶段聚类显示最优分类数为4类;四分类系统聚类结果显示,97.8%的对象为2类;三分类系统聚类结果显示,98.5%的对象为1类;四分类的K-means聚类结果显示,14.0%、16.0%、30.6%和39.5%的对象被分别分为1~4类;三分类聚类结果显示,26.3%、34.2%和39.5%的对象被分别分为1~3类。结论集中式供水风险评价适宜分为分高、较高、较低、低四类;K-means聚类方法更适宜应用于该领域。
[Abstract]:Objective to explore a risk assessment method suitable for the characteristics of centralized water supply. Methods 1 302 centralized water supply facilities in China were selected and their risk index was calculated. Two-stage clustering is used to calculate the optimal classification number. Two methods, systematic clustering and K-means clustering, were used to analyze the risk index of the study object by using the optimal classification number and the standard recommended three classification methods. Results the optimal classification number of two-stage clustering was 4, the result of four-classification system showed that 97.8% of objects were classified into two classes, and the result of three-stage clustering showed that 98.5% of objects were classified as 1. The K-means clustering results of the four classifications showed that 30.6% and 39.5% of the objects were classified into 1 and 4 categories, respectively, and the results of the three classification clustering showed that 34.2% and 39.5% of the objects were classified into three groups, respectively. Conclusion the risk assessment of centralized water supply can be divided into four categories: high, low and low, and K-means clustering method is more suitable for this field.
【作者单位】：中国疾病预防控制中心农村改水技术指导中心;重庆医科大学公共卫生与管理学院;
【基金】：2013年国家卫生计生委卫生公益性行业科研专项(201302004)
【分类号】：R123.1

【参考文献】