在依据空测环境数据通过模糊聚类分析确定空中环境谱区域划分中,提出了对分类的各类数据总体分布齐次性检验加以判别和归类的方法。通过对典型应用实例中3种检验方法(t分布假设检验、秩和检验、斯米尔诺夫检验)检验结果的分析,确定使用秩和检验与斯米尔诺夫检验相结合的方法进行空中环境谱模糊聚类分析的假设检验,使空中环境谱的区域划分更为科学合理。
Air environmental spectrum is the important composing part of airplane service environmental spectrum. Large amount of spectrum data without evident dividing line is classified by fuzzy clustering method. The homogeneity between their parent distribution functions is tested by three commonly used hypothesis testing methods: t distribution hypothesis test, rank-sum test and Smirnov test. According to their characteristics and the results of air environmental spectrum clustering analysis, the t distribution hypothesis testing is more sensitive than the other two and does not meet the requirements of air environmental spectrum compiling. As a result, the latter two methods are associated and applied to the hypothesis testing of air environmental spectrum clustering analysis.