cvanmf.models

Load existing Enterosignature models.

Attributes

FIVE_ES_COLORS
logger

Classes

Signatures

Definition of an existing signature model.

Functions

five_es(→ Signatures)

The 5 Enterosignature model of Frioux et al.

Module Contents

class cvanmf.models.Signatures

Bases: NamedTuple

Definition of an existing signature model.

This provides the definition of existing signatures required to reapply the signature model to new data. Where Decomposition stores the input and H matrix, these are not necessary for transforming new data. Rather, we only need the W matrix, the colors associated with each signature (for consistency of representation), and the preprocessing steps (to match features in the new data with those in the W matrix).

reapply(y: pandas.DataFrame, **kwargs) → Decomposition

Transform new data using this signature model.

Parameters:

y – New data of same type as the existing model.

citation: str | None = None

Citation when using this model.

colors: List[str] | Dict[str, str]

Color for each signature in the model.

feature_match: FeatureMatch

Function to map features in new data to those in the model W matrix.

input_validation: InputValidation

Function to validate and potentially transform input table. Defaults to identity function

w: pandas.DataFrame

Feature weights (W matrix) for this model.

cvanmf.models.five_es() → Signatures

The 5 Enterosignature model of Frioux et al. (2023, https://doi.org/10.1016/j.chom.2023.05.024). A summary of this model can also be found on the website https://enterosignatures.quadram.ac.uk. The reapply method for this model will normalise (total-sum-scale) input data after applying filters to match model format, so data provided does not need to be normalised.

Returns:

5 Enterosignature model

Type:

Signatures

cvanmf.models.FIVE_ES_COLORS

cvanmf.models.logger: logging.Logger