Hi to all
On the next screen shot an indicator that found days a go…
The question is if it makes any sense marking all the bars according to the color and submit the data base to a multiclass clasification model (DNN, XGBosst, Random Forest, etc…) and get the predictions for the next bar and Open or close trades according to the predictions?
What makes me think is that I wont have an imbalance situation, that every single bar will have a label making the model very robust in terms of information.
Or by labeling all the data what I will do is confuse the algorithm as the features change and the labels wont and get to no where?
Ani ideas or coments more than apreciate
