Handling Two Directional Outputs in Deep Learning Trading Models
Summary
This exchange asks how to make a deep-learning model produce two predictions for futures trading, one for a long view and one for a short view. The response explains that producing multiple outputs for a regression or classification model may require a custom loss function, depending on the model setup.
As a simpler alternative, it suggests predicting a return and then using that predicted return to make a long-or-short decision. Another proposed framing is to classify outcomes by whether a prediction crosses a chosen threshold. The discussion offers no implementation details, performance evidence, or guidance for selecting and validating the threshold, so it describes modeling options rather than a tested trading method.
Key ideas
- A model with multiple outputs may require a custom loss function.
- A predicted return can be converted into a long-or-short trading decision.
- Threshold-based classification is another way to represent directional forecasts.
- The exchange provides no empirical results or threshold validation method.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.