Consequent Parameter Estimator¶
-
class
pyfume.EstimateConsequentParameters.ConsequentEstimator(x_train, y_train, firing_strengths)¶ Bases:
objectCreates a new consequent estimator object.
- Parameters
x_train – The input data.
y_train – The output data (true label/golden standard).
firing_strengths – Matrix containing the degree to which each rule fires for each data instance.
-
suglms(global_fit=False, df=0)¶ Estimates the consequent parameters in the first-order Sugeno-Takagi model using least squares.
- Parameters
global_fit – Use the local (global_fit=False) or global (global_fit=True) least mean squares estimates. Global estimates functionality is still in beta mode, so use with caution.
df – default value returned when the sum of grades equals to one (default = 0).
- Returns
The parameters for the consequent function.
-
zero_order()¶ Estimates the consequent parameters of the zero-order Sugeno-Takagi model using normalized means.
- Parameters
df – default value returned when the sum of grades equals to one (default = 0).
- Returns
The parameters for the consequent function.