ConstantPrior

ConstantPrior

The Constant prior.

Returns a single constant value for every input point: f(x) = c. Can be fitted to the mean of the training outputs.

Constructor

new ConstantPrior(c)

Source:
Parameters:
Name Type Description
c number

the constant prior value

Classes

ConstantPrior

Members

c :number

Source:
Type:
  • number

Methods

fit(_inputs, outputs)

Description:

  • Fits the prior to the arithmetic mean of outputs. Matches Rust fit(_, training_outputs) { self.c = training_outputs.mean() }. nalgebra's mean() divides by N (element count), matching JS arithmetic mean.

Source:
Parameters:
Name Type Description
_inputs Array.<Array.<number>>

ignored
outputs Array.<number>

training output values

prior(inputs) → {Array.<number>}

Description:

  • Returns a constant vector of length inputs.length, each entry = this.c. Matches Rust prior(input) → DVector::from_element(input.nrows(), self.c).

Source:
Parameters:
Name Type Description
inputs Array.<Array.<number>>

one sample per row
Returns:

constant array, length = inputs.length

Type
Array.<number>

(static) default(_inputDimension) → {ConstantPrior}

Description:

  • Creates the default ConstantPrior with c = 0. Matches Rust ConstantPrior::default(_input_dimension) → Self::new(0f64).

Source:
Parameters:
Name Type Description
_inputDimension number

ignored
Returns:
Type
ConstantPrior