Simpful Versions Save

• Introduced support for categorical variables in consequent evaluation with template replacement.
• Changed license to AFL v3.0

Fixed a bug with matplotlib version >=3.8

You can now replace fuzzy rules in your rule base with the replace_rule method, and visualize or export the current rule base with get_rules.

You can now plot membership values of a given element of the universe of discourse using the plotting facilities of simpful. Just pass the element argument in the draw, plot, and plot_variable methods, or a 'element_dict' to the produce_figure method. The argument 'TGT' has been kept for retro-compatibility, but it is now considered deprecated and it will be removed in future releases.

Thanks to some valuable contributions from @akdenizince, we now support weighted fuzzy rules and extended aggregation functions for Mamdani inference. A pre-implemented probabilistic OR (probor) function was added to be used as aggregation function, along with Python built-ins (sum and max). If you have special aggregation functions, you can also pass them via the aggregation_function argument (code here, documentation here).

A pointer to an aggregation function can now be specified when calling the Mamdani_inference method, with the argument aggregation_function.

This new version supports weighted fuzzy rule bases in Takagi-Sugeno fuzzy systems.

Simpful can now parse any alphanumeric character and the _ .