Metadata-Version: 2.1
Name: pyanfis
Version: 0.1
Summary: An pytorch-based package to use ANFIS for AI
Author-email: Vicente Feced Mas <vicente.feced.mas@gmail.com>
Project-URL: Homepage, https://github.com/VicenteFecedMas/pyanfis
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE

# pyANFIS

Welcome to pyANFIS! here you will be able to find a framework that will allow you to use **Fuzzy Logic** with usual pytorch layers.
This framework is based on [Jang's](https://www.researchgate.net/publication/3113825_ANFIS_Adaptive-Network-based_Fuzzy_Inference_System?enrichId=rgreq-15825cac70a3ae795992310484420cab-XXX&enrichSource=Y292ZXJQYWdlOzMxMTM4MjU7QVM6MTU5MDc1MDY1MTQ3MzkyQDE0MTQ5Mzc4NTk3MzI%3D&el=1_x_2&_esc=publicationCoverPdf) original paper, although it is going to implement several more things (listed below).

## 2024 Roadmap

- [x] Jang's Original ANFIS.
- [ ] Create documentation for each class and function.
- [ ] Create an installation tutorial.
- [x] Consequent parameters can be estimated using backpropagation.
- [X] Type 1 (Tsukamoto) consequents can be used.
- [ ] Type 2 (Lee) consequents can be used.
- [ ] Functions inside a universe can merge when 2 functions cover a similar area.
- [ ] Automatically not use rules when they are not relevant.
- [X] Display what rules have been fired with a certain data.
- [ ] Create a method to substitute a trained ANFIS with a surface like in [Matlab](https://www.mathworks.com/help/fuzzy/genfis.html#d126e35957).
