semasuka / credit-card-approval-prediction-classification Goto Github PK

Hi,

I think the statement about bear bull market is wrong.

If default customer is labeled as 1 in the response label:

I think under Bull market, precision is more important. Since you want customers "that are classified as bad by your model" to be all genuine bad customers, aka, you allow some bad customer to be passed.

In contrast, under Bear market, recall is more important. You want customers "that are not classified as bad by your model" to be all genuine good customers, aka, you allow some good customers to be wrongly classified as bad.

Best,

While clicking on Prediction I am getting an error.

Secrets file not found. Expected at: workspace/Credit-card-approval-prediction-classification/.streamlit/secrets.toml