Community Tip - You can change your system assigned username to something more personal in your community settings. X
Good morning all,
how can I change the dataset in a model without recreate model and its references?
I would exchange the full dataset of a completed model, but i don't want to recreate it.
Many thanks in advance
Solved! Go to Solution.
Good Morning @cmorfin,
Yes, that's all clear.
First, I don't want to upload additional data, because we usually take up data from the beginning without time filters. So, in this way I have to check the start data to not duplicate data.
Second, when I'm going to retrain the model It will be created a new Model Id and I've to set again in the property of my parallel prediction thing this new value.
Anyway, I will do as the usual scenario.
Many Thanks
Regards
Is it possible that you clarify what exactly you need to do ?
Do you want to retrain the model with a new dataset ?
or do you want to do something different ?
Thank you
Christophe
Hi @cmorfin,
I would create a new model exchanging the dataset.
By retraining, I cannot select a new dataset and it will be creating a new model Id.
Uploading additional data in data section, I cannot truncate existing dataset.
Is possible to create a new model from existing model, selecting a new dataset?
Otherwise I have to recreate a new model from zero, with new dataset.
Many Thanks
The short answer is no we cannot do this.
I though do not quite understand why you want to discard the original dataset.
The usual scenario is to create a model with dataset1.
Then it can happen that we get a new dataset2 that reflect a more recent behaviour of the system, so we can upload new data to dataset1, then retrain the model.
Indeed we cannot exclude the previous data from dataset1 but since they are also part of the former behaviour this is usually something that the model needs to be aware of, so it can handle previous and current situation.
Is there a specific reason why you do not want to keep dataset1 ?
Also why is getting a new model id an issue for you ?
Thank you
Kind regards
Christophe
Good Morning @cmorfin,
Yes, that's all clear.
First, I don't want to upload additional data, because we usually take up data from the beginning without time filters. So, in this way I have to check the start data to not duplicate data.
Second, when I'm going to retrain the model It will be created a new Model Id and I've to set again in the property of my parallel prediction thing this new value.
Anyway, I will do as the usual scenario.
Many Thanks
Regards