balanced test set

Comments

• EJ

  October 2023

  I've asked the team
• koenderks

  October 2023 edited October 2023

  Hi may01dz,

  You cannot do this in the machine learning module directly but requires the manual inspection of the data and specification of a variable in the data set that indicates which observations belong to the test set (indicated by 1, you have manually balanced these across classes) and which observations belong to the training set (indicated by 0). You can then use this balanced test set for training the model.

  Note that a balanced test set is an ideal scenario, but is often impractical since one class may occur in the data naturally much more often (e.g., sick vs. healthy people, fraud vs non-fraud).

  Hope that clarifies a little!
• may01dz

  October 2023

  I believe I have grasped your instructions and have executed the following steps:

  I duplicated the data contained within the 'Churn' column and pasted it into a new column which I designated as 'testIndicator.' I have then proceeded to select a representative sample comprising ten percent of customers classified as 'NO,' and an additional ten percent classified as 'YES.' I have represented these selections within the Colum as '1', whereas the remaining eighty percent of customers were represented as '0.'

  telcoCustomerChurn.csv

  I kindly request your confirmation on the correctness of my approach. The file is attached for your reference.

  

  
  
• koenderks

  October 2023

  Yes, this looks good! You now have a balanced test set and you can use the 'testIndicator' variable as the 'Test set indicator' (as shown below) to index the test set when training any of the algorithms in the machine learning module!

  

  
  
• may01dz

  October 2023

  Thanks, that's what I did.