python-fuzzy-cognitve-maps
python-fuzzy-cognitve-maps copied to clipboard
Using fuzzy cognitive maps for multivariate data forecasting in Python 3.8.
Using Fuzzy Cognitive Maps For Multivariate Data Forecasting
In a nutshell
Fuzzy cognitive maps can be used as a simple model for multivariate data forecasting. Reacreation of this approach in Python 3.8 yields interesting results, albeit weaker than expected. This method is tested against the UWave time series dataset.
How to use the repository
-
dataset_instructions.txt
- explains how to set up the UWave dataset. -
requirements.txt
- use for Python pip install.
Key project files
-
fcm_train.py
- trains the model and stores it in a json file. -
fcm_test.py
- tests the models given in a json file. -
fcm.py
- shortcut for doing both of the above.
Branches
All the additional branches in the repository are different configurations of the model used for experiments. Most of these have been used in the project report, which you may read below or on Overleaf.
The project report