-
Notifications
You must be signed in to change notification settings - Fork 0
guillermo-arce/stock_market_models
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
Welcome to the README, please find below some indications. Directory structure explained: - data: All the raw datasets for the different models. They are picked by "data_processing.py" in order to be pre-processed. - pytorch: Everything related with the pytorch models (except the raw dataset that is in "data" directory) ---- trained_models: Contains the already trained models, ready to be loaded in "torch_model.py" to make predictions. ---- torch_model.py: The script for creating models, training/loading them and make predictions. - tensorflow: Everything related with the tensorflow models (except the raw dataset that is in "data" directory) ---- trained_models: Contains the already trained models, ready to be loaded in "tf_model.py" to make predictions. ---- tf_model.py: The script for creating models, training/loading them and make predictions. - data_processing.py: Script for pre-processing raw data. It is needed to pre-process data before making predictions with the specific model scripts ("tf_model.py" or "torch_model.py") - ta_functions.py: Auxiliary script for "data_preprocessing.py". It helps with the technical indicators addition. HOW-TO: 1º Open "data_preprocessing.py" (Spyder IDE is recommended). 2º Execute "data_preprocessing.py" with the desired parameters, just following the script indications. 3º At this point, we can find the pre-processed data and the scaler in "tensorflow" or "pytorch" according to what we have chosen in the "data_preprocessing.py" script. 4º Open "tf_model.py" or "torch_model.py", depending on which data you have pre-processed (Spyder IDE is recommended). 5º Execute it, consistently with parameters specified in "data_preprocessing.py". Again, following the scripts indications will be helpful. 6º Repeat as desired with the different possibilities: TensorFlow with 100/1 or 100/10 models, PyTorch with sept/oct model, PyTorch with oct/nov model and PyTorch with nov/dec model.
About
TensorFlow and PyTorch models. LSTM neural networks in order to make predictions on (time series) financial data are created with both libraries.
Topics
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published