[英]Hyper-parameter Tuning for a machine learning model
Why a hyper-parameter like regularization parameter (a real number) cannot be trained over training data along with model parameters?为什么像正则化参数(实数)这样的超参数不能与 model 参数一起在训练数据上进行训练? What will go wrong?
go 会出现什么问题?
This is generally done to prevent overfitting.通常这样做是为了防止过度拟合。 Model parameters are trained using the training set.
使用训练集训练 Model 参数。 Hyper-parameter tuning is done using a validation set that is (ideally) completely independent of the training data.
超参数调整是使用(理想情况下)完全独立于训练数据的验证集完成的。 The final performance should be evaluated on a test set.
最终性能应在测试集上进行评估。 Typical splits are 80/10/10 or 60/20/20.
典型的分割是 80/10/10 或 60/20/20。
If you tune your hypermeters on the training set, you will very likely vastly overfit and suffer a performance hit on the test set.如果你在训练集上调整你的超表,你很可能会严重过度拟合并在测试集上遭受性能打击。
Try it out, See the difference in performance on your test set when you do hyper-parameter tuning on the training set.试试看,当你对训练集进行超参数调优时,看看你的测试集上的性能差异。 vs on a separate validation set
vs 在单独的验证集上
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.