线性回归 model 中 MAPE 的解释？

Question

I made a Linear Regression in Python.I computed the MAPE as follows, and got a value of 0.052:

mean_absolute_percentage_error(y_test, pred_test)

Does this mean that my mean percentage error is 0.052% or 5.2%?

Answer 1

In your case, it means 5.2%. You can easily test this, for example if we have the prediction to be 80% of the true, then the error will be 20% and you can see you get 0.2:

from sklearn.metrics import mean_absolute_percentage_error
y_true = [3.3, 1.5, 2.1, 7.2]
y_pred = [0.8*i for i in y_true]

mean_absolute_percentage_error(y_true, y_pred)
Out[9]: 0.19999999999999996