MSE（Mean Squared Error）是均方误差，又称均方差，是n个数据预测结果误差平方的均值，是线性回归模型最常用的损失函数。

A.The attribute values and classlabels may suffer from stochasticnoise，systematic noise，and random artefacts.
B.The quality of the induced classifier depends ontraining examples.
C.Apart from low error rate，some applications require that the classifer provides the reasons behind the classification.
D.The simplest performance criterion is rmse.