python-pandas-dataframe-data填充多维统计信息
我有一个数据框架，其中列用于说明恒星的不同特征，行用于说明不同恒星的测量。（类似这样的） \属性uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu star1 star2 star3 在某些测量中，特定属性的误差为-1.00，这意味着测量有误 在这种情况下，我想放弃测量 一种方法是删除整行（以及错误不是-1.00的其他属性） 我认为可以使用基于所有其他度量的分布生成的值来填充错误度量，也就是说，给定其他良好的属性，该属性应该具有该值，以减少整个数据集的错误 我指的这个想法有一个合适的名字吗？ 您将如何应用这种算法 我是一个单独项目的学生，所以我非常希望能给出详细的理论答案（： 编辑 进一步阅读后，我认为我所指的是回归插补 所以我想我的问题是-我如何以最有效的方式在数据框架中实现多维线性回归python-pandas-dataframe-data填充多维统计信息,pandas,machine-learning,statistics,padding,least-squares,Pandas,Machine Learning,Statistics,Padding,Least Squares,我有一个数据框架，其中列用于说明恒星的不同特征，行用于说明不同恒星的测量。（类似这样的） \属性uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu star1 star2 star3 在某些测量中，特定属性的误差为-1.00，
谢谢！这项技术被称为多重插补（）。这个问题更适合交叉验证。如果你问这个问题，你就不能自己动手了：