HistogramRepresentation

class mlquantify.representations.HistogramRepresentation(bins=(10,), range=(0.0, 1.0), mode='histogram', features=None, partition_blocks=False, bin_edges='fixed', laplace_smoothing=False)

Histogram-based representation.

This representation computes histograms for each feature (or selected subset) independently, and concatenates the bin frequencies into a single vector. When partition_blocks=True, the attribute block_slices_ is populated with the slice objects that identify each bin group in the concatenated output.

Parameters:

binsint or array-like of shape (n_features,), default=(10,)

The number of bins for each feature. If an integer, the same number of bins is used for all features.

rangetuple of shape (2,), default=(0.0, 1.0)

The lower and upper bounds for the histogram range.

modestr, default=”histogram”

The mode of the histogram. Options are: - “histogram”: Compute the histogram counts. - “onehot”: Compute a one-hot encoding of the histogram. In this mode, the output is a binary vector indicating which bin each value falls into, averaged over all samples.

featuresarray-like of shape (n_features,), default=None

The indices of the features to use. If None, all features are used to compute the histogram representation.

partition_blocksbool, default=False

Whether to partition the output into contiguous blocks corresponding to each feature’s bins. When partition_blocks=True, the returned representation is still a single 1-D concatenated vector, but it is organized in feature-wise blocks: all bins for the first selected feature appear first, then all bins for the second selected feature, etc. In this case the attribute block_slices_ is populated with a tuple of slice objects that identify the start/stop indices for each block, allowing easy extraction of per-feature bin groups from the concatenated vector. When partition_blocks=False, the same concatenation is returned but no block_slices_ metadata is stored.

Examples

>>> from mlquantify.representations._histogram import HistogramRepresentation
>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> scores = rng.uniform(0, 1, (200, 1))
>>> y = (scores[:, 0] > 0.5).astype(int)
>>> rep = HistogramRepresentation(bins=(8,)).fit(scores, y)
>>> rep.transform(scores[:10]).shape
(8,)

fit(X, y, classes=None, sample_weight=None)

Fit the representation to labelled training data.

Validates shapes, stores the class labels, delegates internal fitting to _fit, and verifies that the subclass set class_representations_ during that call.

Parameters:

Xarray-like of shape (n_samples, n_features) or (n_samples,)

Feature matrix or pre-computed score array for the training instances.

yarray-like of shape (n_samples,)

Class labels for each training instance.

classesarray-like of shape (n_classes,) or None, default=None

Explicit list of class labels. If None, the unique values in y are used.

sample_weightarray-like of shape (n_samples,) or None, default=None

Per-sample weights forwarded to _fit.

Returns:

selfBaseRepresentation

The fitted representation object.

Raises:

ValueError

If X and y have inconsistent lengths or X is zero-dimensional.

AttributeError

If the subclass did not define class_representations_ inside _fit.

Examples

>>> from mlquantify.representations import HistogramRepresentation
>>> import numpy as np
>>> X = np.random.default_rng(0).uniform(0, 1, (100, 1))
>>> y = (X[:, 0] > 0.5).astype(int)
>>> rep = HistogramRepresentation(bins=(5,)).fit(X, y)
>>> rep.class_representations_.shape
(2, 5)

transform(X)

Compute the histogram representation for a set of instances.

Each feature (or selected subset) is binned independently, and the bin frequencies are concatenated into a single vector. When partition_blocks=True, the attribute block_slices_ is populated with the slice objects that identify each bin group in the concatenated output.

Parameters:

Xarray-like of shape (n_samples, n_features) or (n_samples,)

Feature matrix or 1-D score array.

Returns:

representationndarray of shape (n_bins_total,)

Normalized histogram vector (sums to 1 per feature-bin group).

Examples

>>> from mlquantify.representations._histogram import HistogramRepresentation
>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> scores = rng.uniform(0, 1, (200, 1))
>>> y = (scores[:, 0] > 0.5).astype(int)
>>> rep = HistogramRepresentation(bins=(8,)).fit(scores, y)
>>> rep.transform(scores[:10]).shape
(8,)