I have a working implementation of a multivariate, multistep time series prediction using LSTMs on stock ticker data (OHLC). I wanted to extend this by adding some additional features (Technical Indicators) and scaling the input data as the new features are not of the same magnitude as the price. I can add the features and scale the input data, train the model and make predictions but I'm struggling to invert the scaling. Below is the github address with the problem laid out in a Jupyter Notebook. I need someone to get the scale inversion working for me on BOTH a single prediction (for e.g X_test, y_test) and on the entire test set in a single step (for eg. X_test, y_test).
Github: [login to view URL]
6 freelancers are bidding on average $226 for this job
hello I am mashud, expert in deep learning and keras. I will complete this project in 1 day perfectly. please contact me to discuss more about the project. thanks.