让我们假设我们有一个数据集，它大概是Therefore we have a variation of 20% of the dataset. My first idea was to use the