基于模糊粗糙集的气象因子与森林火灾相关性研究

Correlation between meteorological factors and forest fire based on fuzzy rough sets method

为了揭示各气象因子对森林火灾发生的影响程度,随机抽取了湖南省1 041场森林火灾历史数据,收集了森林火灾发生时的温度、相对湿度、风速及火灾发生前24 h降水量等4个因子数据,将其标准化并进行K-均值聚类,再用自然断点法处理后得到属性决策表。各气象因子采用相应三角隶属函数模糊化后,确定了森林火灾气象因子模糊粗糙集模型,计算各气象因子与决策属性的依赖度,对其进行归一化后,得到温度、相对湿度、风速、降水量对森林火灾发生的影响值依次为:0.116 2、0.265 0、0.212 4、0.406 3。定义了灾均比(ZJB)的概念,灾均比分析结果印证了研究结果的正确性。

In order to reveal the influence degree of different meteorological factors on forest fire, the historical fire data of one thousand and forty-one times forestifres were randomly sampled in Hunan Province, and four kinds of factor data （temperature, relative humidity, wind speed and 24 hours precipitation） were collected, then the date were standardized and were clustered with K-means clustering method, andifnally the date were processed by adopting the natural breakpoint method and the attribute decision-making table were obtained; The after the various meteorological factors were treated with fuzziifcation by adopting the corresponding triangular membership functions, the fuzzy rough set models of forestifre meteorological factors were determined; Then the dependence of each meteorological factors and decision attributes were calculated, and the dependence values were treated with normalized processing method,ifnally, the effective values of temperature, relative humidity, wind speed and 24 hour precipitation were gotten, they are 0.1162, 0.2650, 0.2124, 0.4063. According the results, the concept of ZJB was deifned, the ZJB analysis results conifrmed the correctness of the conclusion.