First version

First version released in PyPi. The implementation of Data drive fuzzy cognitive map paper with some other features not published yet.

This version includes:

• FCM definition
• Map inference using Kosko or Papageorgius functions. Also, support the following activation functions:
  • biestate
  • threestate
  • saturation
  • sigmoid
  • sigmoid_hip
  • relm (Rectified Linear Unit)
  • fuzzy (Only for fuzzy concepts)
  • gceq (Greater conditional equality)
  • lceq (Lesser conditional equality)
• Map decision using Mean, Last or Excited functions
• The vectorized implementation of the inference process
• Map topology generation for supervised classification problems
• Support discrete and continuous features
• Transform continuous features to fuzzy concepts with the algorithms:
  • cmeans-gap (cmeans with a proposed clusters estimation method)
  • cmeans-sil (cmeans with fuzzy silhouette)
  • cmeans-cfe (cmeans with CFE)
  • rl-cmeans (robust learning fuzzy cmeans implementation)
• Support several relation functions between concepts:
  • supp (support)
  • conf (connfidence)
  • lift
  • ods (ods ratio)
  • bayes (probabilistic definition)
  • pos_inf (positive influence)
• Scikit-learn's estimators behavior in the generation process, to make easy any test or experimentation