我认为你最好的选择是忽略异常值 - 即丢弃前 10% 和后 10%，然后计算平均值
或者，也许重新思考你想要做什么，也许有某种方法为选民分配权重
I think your best bet is to ignore the outliers - ie discard the top 10% and the bottom 10% and then calculate the average
Alternatively, maybe re-think exactly what you want to do, maybe have some way of assigning weights to the voters
我在网上听说中位数是平均值的常用替代品，通常在倾斜的设置中。在对称数据集中，均值、中位数和众数相同。
I've heard online that median is a commonly used alternative to mean, usually in skewed settings. In symetrical data sets, the mean, median and mode are the same.
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
暂无简介
文章 0 评论 0
接受
发布评论
评论（2）
我认为你最好的选择是忽略异常值 - 即丢弃前 10% 和后 10%，然后计算平均值
或者，也许重新思考你想要做什么，也许有某种方法为选民分配权重
I think your best bet is to ignore the outliers - ie discard the top 10% and the bottom 10% and then calculate the average
Alternatively, maybe re-think exactly what you want to do, maybe have some way of assigning weights to the voters
我在网上听说中位数是平均值的常用替代品，通常在倾斜的设置中。在对称数据集中，均值、中位数和众数相同。
I've heard online that median is a commonly used alternative to mean, usually in skewed settings. In symetrical data sets, the mean, median and mode are the same.