decisionintelligence
Using multiple variables to predict one of the single variables How to handle the data set？
In your examples, both multivariable and univariable, what should I do with the data? All I can think of so far is to add a linear layer after calling the model and turn multiple channels into one channel, what about the label?
Hello, Thank you for your attention and questions. Sorry, our code currently cannot directly predict single variables with multiple variables. But we can implement this feature by modifying our framework, and we also believe that this is something that should be considered for the future. We have already adjusted its priority to a higher level. But recently, due to rushing to ICLR, I may not have time to complete it. We will complete the adaptation as quickly as possible.
Hello, now we support this function：https://github.com/decisionintelligence/TFB/blob/master/docs/tutorials/steps_to_predict_only_a_subset_of_input_variables.md