关键词: China Haze K-means clustering algorithm Natural breakpoint method Perception Potential Conflict Index (PCI)

Mesh : Aged Air Pollution / analysis Child China Conservation of Natural Resources Humans

来  源:   DOI:10.1007/s11356-021-17384-8

Abstract:
The concept of haze habituation was proposed based on haze perception and behavior in this paper. This study employed factor analysis and Potential Conflict Index (PCI) to analyze the dimensions, degrees, and internal differences of the public\'s haze habituation. Then, K-means clustering algorithm was applied to classify the public into four categories. The entropy method was used to quantitatively evaluate the public\'s haze habituation, and the natural breakpoint method was used to grade it into five levels. Finally, an ordered logistic regression model was chosen to analyze the influencing factors of the public\'s haze habituation. The results indicate that: (1) The public\'s haze habituation can be measured from five dimensions: protective behavior, haze reduction behavior, haze attention, life impact perception, and health impact perception. The public had the same views on protective behavior, haze reduction behavior, life impact perception, and health impact perception. However, there is a wide divergence among the public on the haze attention; (2) Based on the above five dimensions, the public can be divided into the protective sensitive group, attention sensitive group, health sensitive group, and environmental protection sensitive group; (3) Generally, the public has a low haze habituation where the protective behavior, haze reduction behavior, and health impact perception are the crucial elements; (4) Gender, self-health assessment, and travel mode have a significant positive impact on the public\'s haze habituation, respectively. Age, the family with elders or children, and annual family income have a significant negative impact on the public\'s haze habituation, respectively.
摘要:
本文提出了基于雾霾感知和行为的雾霾习惯概念。本研究采用因子分析和潜在冲突指数(PCI)来分析维度,度,以及公众雾霾习惯的内部差异。然后,应用K-means聚类算法将公众分为四类。采用熵值法对公众的雾霾习惯进行了定量评价,并采用自然断点法将其分级为5级。最后,采用有序logistic回归模型对公众雾霾习惯的影响因素进行分析。结果表明:(1)公众的雾霾习惯可以从五个维度进行测量:防护行为,减霾行为,雾霾注意,生活影响感知,和健康影响感知。公众对保护行为也有同样的看法,减霾行为,生活影响感知,和健康影响感知。然而,公众对雾霾的关注存在较大分歧;(2)基于以上五个维度,公众可以分为保护性敏感群体,注意敏感组,健康敏感群体,和环保敏感群体;(3)一般情况下,公众有低雾霾习惯,保护行为,减霾行为,和健康影响感知是关键因素;(4)性别,自我健康评估,出行方式对公众的雾霾习惯有显著的正向影响,分别。年龄,有长辈或孩子的家庭,和家庭年收入对公众的雾霾习惯有显著的负面影响,分别。
公众号