Information for : Delayed biological recovery from extinctions throughout the fossil record
Information for : Delayed biological recovery from extinctions throughout the fossil record
复制标题
信息：整个化石记录中生物从灭绝中恢复延迟
DOI：
--
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
J. Kirchner
中科院分区：
文献类型：
--
作者：
J. Kirchner
与经典的统计显着性估计相比，随机化方法具有两个重要优势。首先，人们不需要能够从数学上推导出零假设下可能结果的分布，因为人们可以通过模拟来构建它。其次，原始数据不需要服从任何特定的分布，因为它们的任何潜在的麻烦特征（偏斜，离群值等）也可以合并到随机化数据集中。在我们的例子中，这很容易通过重新洗牌原始数据值本身的时间序列来生成随机数据集来实现。这具有消除随机化数据中灭绝和起源之间的任何交叉相关性（除了偶然出现的）的效果，同时保留灭绝和起源数据本身的原始分布。
Randomization methods offer two important advantages over classical estimates of statistical significance. First, one need not be able to mathematically derive the distribution of possible outcomes under the null hypothesis, since one can construct it by simulation instead. Second, the original data need not obey any particular distribution, since any of their potentially troublesome features (skew, outliers, etc.) can be incorporated in the randomized data sets as well. In our case, this is easily accomplished by generating the randomized data sets by reshuffling the time-series sequence of the original data values themselves. This has the effect of eliminating any cross-correlation between extinctions and originations (except that which arises by chance) in the randomized data, while retaining the original distributions of the extinction and origination data themselves.