Closed zhouzyhfut closed 4 years ago
Hi,
The outcome in the dataset is a score on the PHQ8-scale (I think it is phq8_labels
= [0 to 24]).
The labels in which the classifier is trained with are normalized by the max(phq8_labels
). So during training your ground truth (true
) and predictions (pred
) will be between 0 and 1.
true = [1, 0, 0, 1, 0, 1]
pred = [0.8, 0.2, 0.1, 0.7, 0.3, 0.9]
To calculate MAE and RMSE, you will need to rescale your true
and pred
values back to the PHQ-8 scale.
true = max(phq8_labels) * true
pred = max(phq8_labels) * pred
Now you can calculate MAE (mean absolute error) and RMSE (root mean square error), as follows.
MAE = np.mean(np.abs(true - pred))
RMSE = np.sqrt(np.mean(np.abs(true - pred) * np.abs(true - pred)))
I hope that clarifies it.
Ok, i understand your explanation. Thank you for your help.
Hi, i have read your paper. I am doubt that why the metrics (mae,rmse) value in your experiment >1? I am using the mae metric in Keras to train my model, but its value is between 0 and 1? Could you tell me how do you calculate the metrics in your paper?