Hello!
Had some thoughts about features that could help improve finding/developing a successful strategy.
1. Weighted F-Score from "Financial Markets Prediction with Deep Learning" by Jia Wang et al.
2. Predicted Profit and Loss from an entry signal. How would this be done... I don't know. But on the user side we could set a min and max for Profit and Loss to happen within Bars To Target. The model would be created with the entry signals and the associated profit and loss for that entry. The predictions would trigger an entry signal and set their profit as TakeProfit and loss as StopLoss.
3. Allow entry triggers for model creation. Enabled via a checkbox(?). For example, the model creation would train for entries from the moment CrossAbove(MACD(12,24,9),0,1). This could help direct model creation to fit the strategy we are developing. Could this lead to overfitting...maybe? Won't know till we try.
Had some thoughts about features that could help improve finding/developing a successful strategy.
1. Weighted F-Score from "Financial Markets Prediction with Deep Learning" by Jia Wang et al.
2. Predicted Profit and Loss from an entry signal. How would this be done... I don't know. But on the user side we could set a min and max for Profit and Loss to happen within Bars To Target. The model would be created with the entry signals and the associated profit and loss for that entry. The predictions would trigger an entry signal and set their profit as TakeProfit and loss as StopLoss.
3. Allow entry triggers for model creation. Enabled via a checkbox(?). For example, the model creation would train for entries from the moment CrossAbove(MACD(12,24,9),0,1). This could help direct model creation to fit the strategy we are developing. Could this lead to overfitting...maybe? Won't know till we try.
0