I've trained a KNNRegressor on predicting used car prices. A given car's actual selling price is R289,995. My model predicts R260,911. I want to be able to tell the user My knn model predicts the deal is +-10% cheaper than the actual selling price.

To accurately present this to a user and to account for the model's chance of error, would using the mean error percent (14%) in this case, be a decent way to create an approximate range? Or is this not statistically valid.

So would it be more valid to say: Prediction range: R224,000- R297,000

Model Score


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.