adaptive.AverageLearner1D

class adaptive.AverageLearner1D(*args, **kwargs)

Bases: adaptive.learner.learner1D.Learner1D

Learns and predicts a noisy function ‘f:ℝ → ℝ’.

Parameters

• function (callable) – The function to learn. Must take a tuple of (seed, x) and return a real number.

• bounds (pair of reals) – The bounds of the interval on which to learn ‘function’.

• loss_per_interval (callable, optional) – A function that returns the loss for a single interval of the domain. If not provided, then a default is used, which uses the scaled distance in the x-y plane as the loss. See the notes for more details of adaptive.Learner1D for more details.

• delta (float, optional, default 0.2) – This parameter controls the resampling condition. A point is resampled if its uncertainty is larger than delta times the smallest neighboring interval. We strongly recommend 0 < delta <= 1.

• alpha (float (0 < alpha < 1), default 0.005) – The true value of the function at x is within the confidence interval [self.data[x] - self.error[x], self.data[x] + self.error[x]] with probability 1-2*alpha. We recommend to keep alpha=0.005.

• neighbor_sampling (float (0 < neighbor_sampling <= 1), default 0.3) – Each new point is initially sampled at least a (neighbor_sampling*100)% of the average number of samples of its neighbors.

• min_samples (int (min_samples > 0), default 50) – Minimum number of samples at each point x. Each new point is initially sampled at least min_samples times.

• max_samples (int (min_samples < max_samples), default np.inf) – Maximum number of samples at each point x.

• min_error (float (min_error >= 0), default 0) – Minimum size of the confidence intervals. The true value of the function at x is within the confidence interval [self.data[x] - self.error[x], self.data[x] + self.error[x]] with probability 1-2*alpha. If self.error[x] < min_error, then x will not be resampled anymore, i.e., the smallest confidence interval at x is [self.data[x] - min_error, self.data[x] + min_error].

ask(n: int, tell_pending: bool = True) → tuple[typing.List[typing.Tuple[int, typing.Union[float, numpy.float64, int, numpy.int64]]], list[float]]

Return ‘n’ points that are expected to maximally reduce the loss.

data: dict

property min_samples_per_point: int

property nsamples: int

Returns the total number of samples

pending_points: set

plot()

Returns a plot of the evaluated data with error bars (not implemented

for vector functions, i.e., it requires vdim=1).

Returns

plot – holoviews.element.Path` Plot of the evaluated data.

Return type

`holoviews.element.Scatter * holoviews.element.ErrorBars *

tell(seed_x: Tuple[int, Union[float, numpy.float64, int, numpy.int64]], y: Union[float, numpy.float64, int, numpy.int64]) → None

Tell the learner about a single value.

Parameters

• x (A value from the function domain) –

• y (A value from the function image) –

tell_many(xs: List[Tuple[int, Union[float, numpy.float64, int, numpy.int64]]], ys: Sequence[Union[float, numpy.float64, int, numpy.int64]]) → None

Tell the learner about some values.

Parameters

• xs (Iterable of values from the function domain) –

• ys (Iterable of values from the function image) –

tell_many_at_point(x: Union[float, numpy.float64, int, numpy.int64], seed_y_mapping: dict[int, typing.Union[float, numpy.float64, int, numpy.int64]]) → None

Tell the learner about many samples at a certain location x.

Parameters

• x (float) – Value from the function domain.

• seed_y_mapping (Dict[int, Real]) – Dictionary of seed -> y at x.

tell_pending(seed_x: Tuple[int, Union[float, numpy.float64, int, numpy.int64]]) → None

Tell the learner that ‘x’ has been requested such that it’s not suggested again.