Stock prediction using LSTM Machine Learning
Many-to-One LSTM taking price and volume for each minute as inputs and a single heuristic output (measured with future prices). Using Keras library (a wrapper for tensorflow).
-
Models Folder - Generated models from the script
-
RawData Folder - Stock data in csv format
-
Run.py
- Parses data from RawData folder into a list of times, tickers, opens, and volumes during trading hours
- Split data into train and test data
- Generate models for each minute based on data from earlier in day
- Test models with test data
-
Scratch Folder - Contains files used to test previous stock algorithm ideas