Builders¶

anfis_toolbox.builders.ANFISBuilder ¶

ANFISBuilder()

Builder class for creating ANFIS models with intuitive API.

add_input ¶

add_input(
    name: str,
    range_min: float,
    range_max: float,
    n_mfs: int = 3,
    mf_type: str = "gaussian",
    overlap: float = 0.5,
) -> ANFISBuilder

Add an input variable with automatic membership function generation.

Parameters:

Name	Type	Description	Default
`name`	`str`	Name of the input variable	required
`range_min`	`float`	Minimum value of the input range	required
`range_max`	`float`	Maximum value of the input range	required
`n_mfs`	`int`	Number of membership functions (default: 3)	`3`
`mf_type`	`str`	Type of membership functions. Supported: 'gaussian', 'gaussian2', 'triangular', 'trapezoidal', 'bell', 'sigmoidal', 'sshape', 'zshape', 'pi'	`'gaussian'`
`overlap`	`float`	Overlap factor between adjacent MFs (0.0 to 1.0)	`0.5`

Returns:

Type	Description
`ANFISBuilder`	Self for method chaining

add_input_from_data ¶

add_input_from_data(
    name: str,
    data: ArrayLike,
    n_mfs: int = 3,
    mf_type: str = "gaussian",
    overlap: float = 0.5,
    margin: float = 0.1,
    init: str | None = "grid",
    random_state: RandomStateLike = None,
) -> ANFISBuilder

Add an input inferring range_min/range_max from data with a margin.

Parameters:

Name	Type	Description	Default
`name`	`str`	Input name	required
`data`	`ArrayLike`	1D array-like samples for this input	required
`n_mfs`	`int`	Number of membership functions	`3`
`mf_type`	`str`	Membership function type (see add_input)	`'gaussian'`
`overlap`	`float`	Overlap factor between adjacent MFs	`0.5`
`margin`	`float`	Fraction of (max-min) to pad on each side	`0.1`
`init`	`str \| None`	Initialization strategy: "grid" (default), "fcm", "random", or `None`. When `"fcm"`, clusters from the data determine MF centers and widths (supports 'gaussian' and 'bell').	`'grid'`
`random_state`	`RandomStateLike`	Optional seed for deterministic FCM initialization.	`None`

build ¶

build() -> TSKANFIS

Build the ANFIS model with configured parameters.

set_rules ¶

set_rules(
    rules: Sequence[Sequence[int]] | None,
) -> ANFISBuilder

Define an explicit set of fuzzy rules to use when building the model.

Parameters:

Name	Type	Description	Default
`rules`	`Sequence[Sequence[int]] \| None`	Iterable of rules where each rule lists the membership index per input. `None` removes any previously configured custom rules and restores the default Cartesian-product behaviour.	required

Returns:

Type	Description
`ANFISBuilder`	Self for method chaining.