Discussing the article: "Python, ONNX and MetaTrader 5: Creating a RandomForest model with RobustScaler and PolynomialFeatures data preprocessing" - page 2

New comment
 
Yevgeniy Koshtenko #:


It would be interesting to further develop the topic of conveyors and their conversion to ONNX with subsequent use in metatrader. For example, is it possible to add custom transformations to the pipeline and will the ONNX model obtained from such a pipeline be opened in Metatrader? Imho, the topic is worthy of several articles.

 
Why do you need two normalisations?
Why do you even need normalisation for the Forrest?
 
Aleksey Nikolayev #:

It would be interesting to further develop the topic of conveyors and their conversion to ONNX with subsequent use in metatrader. For example, is it possible to add custom transformations to the pipeline and will the ONNX model obtained from such a pipeline be opened in Metatrader? Imho, the topic is worthy of several articles.

Yes, it's a great idea. Perhaps it will be possible to make custom transformations with Function Transform
 
Hello. So I've tried to compile the code but it returned erroes about input_shape and output_shape being parameters conversion not allowed. Did you perhaps ran into the same problem? If so, what could be the correct way?
12
New comment