carmine latch
Hello everyone. I am working on a sample project to build a basic stock prediction app using Dash. The usual approaches that I have come across in making a future prediction is to take the past 60 days of Opening price data for a particular stock, train the model and then make predictions. However I intend to take the Open, High, Low, Close and Adj Close to predict the future stock 'Open' price. Adding these features would result in a 3D array for the training data [samples, bucket of days-60, features]. Since conventional regressors like SVR and RandomForestRegressor only accept 2D data, is there a workaround for this without using Tensorflow/PyTorch? How do I convert my data to 2D without losing any features?
I have also posted this qs in detail here - https://stackoverflow.com/questions/75398540/how-to-model-and-fit-a-multi-dimensional-array-in-svr-and-randomforestregressor
Any help would be appreciated!
viscid thicket