Explain why [50,44,-12] gives 3 (UK) in the web-app sample

[stanciuadrian]

In https://microsoft.github.io/ML-For-Beginners/#/3-Web-App/1-Web-App/README
we test the model with
print(model.predict([[50,44,-12]])) and the authors obtain 3.

I get 1 by running the code.
How is the difference explained?

[Mohammadsiraj07]

I think the difference is most likely due to changes in the training environment rather than an issue with your code.

The prediction isn't hardcoded it's generated by the logistic regression model after training. Even though the tutorial uses
train_test_split(..., random_state=0), you can still get a different prediction if you're using a newer version of the dataset or different versions of libraries such as scikit-learn. Small changes in preprocessing or label encoding can also affect the final model.

If you're using the current dataset and library versions, it's possible to get 1 instead of 3 because the trained model may differ from the one used when the tutorial was written.

It would be helpful if the tutorial mentioned the dataset version and the package versions used when the sample output was created. That would make it much easier for others to reproduce the same result.