Problems about the scale of label in graph-level regression problem #7148
-
|
Hi all! Recently I had a graph-level regression problem. The number of nodes varies from 450 to 140000. I used GIN model with a sum pooling to get the embedding of the graphs. Then I use MLP to regress the labels. The labels vary from 30 to 20000. I would like to ask if there is any need to scale the data to make the labels lie in a narrower range? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
In most cases, you might need a log transform to the labels to normalize the distribution and improve the performance of the regression model. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, I will give it a try. |
Beta Was this translation helpful? Give feedback.
In most cases, you might need a log transform to the labels to normalize the distribution and improve the performance of the regression model.