Credit Card Approval Prediction Using ML on Alteryx

[OOIWJ]

I have no experience in machine learning and only had a 1-week crash course on Alteryx.

I am using 2 datasets found here:

https://www.kaggle.com/datasets/rikdifos/credit-card-approval-prediction

My task is to predict the credit card approval. However, I have one concern, the dataset don't have a column showing whether this customer will be approved/not approved when they apply for credit card. Based on my understanding, during the validation stage we have to use the ML model we created to test the 80% dataset and see whether it predicted correctly. So if we do not have the actual outcome in the dataset, how do we test the accuracy of our model?

Kindly advise on the steps in building a ML model on alteryx, thank you.

[DOLEARY1]

Hi @OOIWJ 

Ideally with a prediction dataset you would have a target to predict.

This dataset requires you to come up with the rules for what a potential approval would look like using analysis of the account since it was opened. There are a lot of details that can be used from monthly payments in the credit record file to see if a payment was on time or late by month. There is also a lot of info on the application record that could be used to inform a decision e.g. employment or age.

Once you have what qualifies as a good bad customer you can then use this as the target for your machine learning model to predict.

If you have no experience with machine learning I'd recommend starting with an easier dataset that contains a target for you to predict pre-defined.