论文部分内容阅读
11.4 极值统计的有关问题11.4.1发生原因不同的统计资料的处理(1)极值的非超值概率的计算如11.1.1节所述,极值统计分析时,必须确定各个数据是独立的,而且它们是属于同一总体的统计量.可是,对自然现象的极值,几乎不可能确定它们是否属于同一总体,所以,较稳妥的做法是按发生原因的不同分别进行极值分析.例如,台风时的暴风和冬季季节风的强风,在风速的绝对值上是有差别的,所以宜按不同的统计资料处理.在难以对每次强风分析其气象原因时,也可以考虑以每年的各月最大值为对象进行解析,将其结果组合起来,再推算所规定的重现期的概率统计量的方法.在把作为研究对象的极值资料按各数据的发生原因或按季、按月分别进行解析时,可以对各不同的组次拟合不同的极值分布函数.为了由该结果对规定的重现期推算其概率统计量,则只要采用卡特(Carter)和恰勒诺(ehallenor)方法,按如下考虑即可.为简单计,首先考虑期间最大值资料,比如已有每年1—12月份各月的最大值资料,由此求出了每月的共12个分布函数的情况.这里,若第j个月的月最大值分布函数以 F_j(x)表
11.4 Related issues of extreme statistics 11.4.1 Processing of statistics with different causes (1) Calculation of non-overvalue probabilities of extremes As described in Section 11.1.1, in extreme statistical analysis, it is necessary to determine that each data is independent However, they are almost impossible to determine whether they belong to the same population for extreme values of natural phenomena, so it is prudent to conduct extreme analysis separately according to the causes , The typhoon storm and the winter season wind strong wind, the absolute value of the wind speed is different, so it should be processed according to different statistics .In each strong wind it is difficult to analyze the weather reasons, but also consider the annual Each month the maximum value for the analysis of the object, the results combined, and then calculate the probability of the specified period of recurrence statistics in the extreme case of the data as the cause of the occurrence of each data or quarterly, according to Month, respectively, different extreme distribution functions can be fitted to different groups.In order to calculate the probability statistics of the results for the specified period of recurrence, it is only necessary to use Carter and Cialer (ehallenor) method, according to the following considerations can be simple for the first time to consider the maximum period of information, such as the maximum monthly data from January to December each year, which obtained a total of 12 distribution function per month Here, if the monthly maximum distribution function of the jth month is represented by F_j (x)