Type Definition lfa::VectorLFA

type VectorLFA<B, O> = LFA<B, Array2<f64>, O>;

Linear function approximator with vector output.

Implementations

impl<B, O> VectorLFA<B, O>

pub fn vector(basis: B, optimiser: O, n_outputs: usize) -> Self where
    B: Space, 

Construct an VectorLFA with a chosen number of outputs using a given basis representation and optimisation routine. The weights are initialised with a matrix of zeros.

pub fn n_outputs(&self) -> usize

Return the dimensionality of the output..

pub fn evaluate<I>(&self, input: I) -> Result<Array1<f64>> where
    B: Basis<I, Value = Features>, 

Evaluate the function approximator for a given input.

pub fn evaluate_index<I>(&self, input: I, index: usize) -> Result<f64> where
    B: Basis<I, Value = Features>, 

Evaluate the ith output of the function approximator for a given input.

pub fn iter<'a, I>(&'a self, input: I) -> OutputIter<'a> where
    B: Basis<I, Value = Features>, 

Iterate sequentially over the outputs of the function approximator for a given input.

Panics if the basis computation fails.

pub fn try_iter<'a, I>(&'a self, input: I) -> Result<OutputIter<'a>> where
    B: Basis<I, Value = Features>, 

Iterate sequentially over the outputs of the function approximator for a given input.

pub fn update<I, E>(&mut self, input: I, errors: E) -> Result<()> where
    B: Basis<I, Value = Features>,
    O: Optimiser,
    E: IntoIterator<Item = f64>, 

Update the function approximator for a given input and error.

pub fn update_index<I>(
    &mut self,
    input: I,
    index: usize,
    error: f64
) -> Result<()> where
    B: Basis<I, Value = Features>,
    O: Optimiser, 

Update the ith output of the function approximator for a given input and error.

pub fn update_with<I>(
    &mut self,
    input: I,
    f: impl Fn(&Features, Array1<f64>) -> Array1<f64>
) -> Result<()> where
    B: Basis<I, Value = Features>,
    O: Optimiser, 

Update the function approximator for a given input, deferring the error computation to a closure provided by the user.