can you guys clarify a bit what is the target normalized by ?
Posted by: ANTONBIRYUKOV2019 @ Oct. 30, 2019, 6:33 a.m.The target in the test set is not normalized. The target in the test set is on the same scale as the as the targets provided to you in the training set (i.e. Viking-Train) and the validation set (i.e. Viking-Validation).
Posted by: untappedenergy2019 @ Oct. 30, 2019, 2:32 p.m.I guess my question is not about the scale (which I could believe since you split datasets on random), but rather the days.In the name of the target one can certainly see the word "Normalized", as well as days in the denominator, hence comes a question of what those days are.
Posted by: ANTONBIRYUKOV2019 @ Oct. 30, 2019, 2:50 p.m.I believe you're referring to the _IP Normalized values. They are a production value that is normalized against # production hours.
If we just include production volumes as reported by the regulator, they will show the total amount of volume that has been produced over a given month. However, a well that comes on early in the month will have a different production profile than one that comes on later int he month. We therefore standardize the production volume against the number of hours that well was producing for over the course of the month to ensure an apples-to-apples comparison.
Posted by: SteveJ @ Oct. 30, 2019, 3:53 p.m.