使用带有分区和样本的SMOTE&交叉验证模型 [英] Using SMOTE with Partition and Sample & Cross Validate Model
问题描述
我想使用"Partition and Sample"和"Partition and Sample"。 (分配给折叠)使用"交叉验证模型"进行k折交叉验证。如何安排模块,使我只在数据的训练部分使用SMOTE,而不是在
数据的其余部分使用?
I want to use "Partition and Sample" (Assign to Folds) with "Cross Validate Model" to do k-fold cross validation. How can I arrange the modules such that I use SMOTE on the training portion of the data only and not on the rest of the data?
如果我使用分层分割在"分区和样本"中我班级的比例仍将保持不平衡。
Also if I use a stratified split in "Partition and Sample" the ratio of my classes will still remain unbalanced.
推荐答案
Hi,
以下是使用SMOTE,Partition和Split模块的一些文档:
Here is some documentation on using the SMOTE, Partition and Split modules:
https://docs.microsoft.com/en -us / azure / machine-learning / studio-module-reference / smote
https://docs.microsoft.com/en- us / azure / machine-learning / studio-module-reference / partition-and-sample
您是否正在尝试构建这样的模型?
Are you trying to build a model like this one?
问候,
Jaya
这篇关于使用带有分区和样本的SMOTE&交叉验证模型的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!