Zipf定律及人工神经网络在环境污染事故预警中的应用研究  

An Applied Study on Law of Zipf and Artificial Neural Network in Warning of Environmental Pollution Accidents

作　　者：袁志忠[1] 周耀渝[1] 曾硕[1] 董哲青[1] 

机构地区：[1]吉首大学生物资源与环境科学学院

出　　处：《上海环境科学》2011年第1期28-30,36,共4页Shanghai Environmental Sciences

基　　金：国家自然科学基金项目资助，编号：30972206

摘　　要：在目前环境问题日益突出的背景下,为研究环境质量变化与经济增长之间的关系,利用齐波夫（Zipf）定律分析了广东省的环境污染事故发生次数与GDP之间的关系,并利用其在图形中规律性幂率区间,结合RBF人工神经网络对未来3年环境污染事故的发生次数进行了预测。结果表明,这2种方法对环境污染与事故预测值均在允许的范围内,平均值分别为6、5、4次。研究结果为有关部门防范未来环境污染事故发生,具有明显的警示作用。To study on the relationship between environmental quality changes and economic growth, the law of Zipf has been applied to analyse the relationship between GDP and the occurrence of environmental pollution accidents in Guangdong Province, whilst its regularity power-law intervals in the graph were used in combination with RBF artificial neural network to predict the occurrence of accidents in the future. The predic- tions by means of those two methods showed that the occurrences were all within permitted extent, i.e., on average 6, 5 and 4 times of accidents might be happening in the next three years, respectively. It＇ could give obvious warning to some departments concerned for preventing from the happening of environmental pollution accidents.

关 键 词：齐波夫定律 人工神经网络 环境污染事故 预警 

分 类 号：X507[环境科学与工程—环境工程]