create_model

Functionality for creating MVDA models like in SIMCA.

creates models

class friendly_mvda.models.create_model.MetaData(primary_obs_id: str | None = None, primary_var_id: int = 0, secondary_obs_ids: list[str] | None = None, secondary_var_ids: list[int] | None = None, batch_id: str | None = None, variable_scaling: list[str] | str = 'none')

Bases: object

batch_id: str | None = None

primary_obs_id: str | None = None

primary_var_id: int = 0

secondary_obs_ids: list[str] | None = None

secondary_var_ids: list[int] | None = None

variable_scaling: list[str] | str = 'none'

friendly_mvda.models.create_model.create_opls_model(data: DataFrame, y: list[str], metadata: MetaData | None = None, qualitative: list[str] | None = None, scaling: str = 'uv') → OplsModel

create a opls model

parameter:

data: data used in the model.

metadata: the metadata.

y: names of the y variables.

qualitative: names of the qualitative variables.

scaling: how the data should be scaled.

returns:

OplsModel: an unfitted OPLS model.

example:

>>> from friendly_mvda.models.create_model import create_opls_model, MetaData
>>> meta_data = MetaData(primary_obs_id="Obs ID", primary_var_id=0)
>>> opls_model = create_opls_model(healthcare_data_index, metadata=meta_data, y=["Test Value"])

will create an opls model with “Test Value” as y variable, Obs ID as primary observation ID, the first row as primary variable ID.

friendly_mvda.models.create_model.create_pca_model(data: DataFrame, metadata: MetaData | None = None) → PcaModel

create a pca model;

parameter:

data: data used in the model.

metadata: the metadata.

returns:

PcaModel: an unfitted PCA model.

example:

>>> from friendly_mvda.models.create_model import create_pca_model, MetaData
>>> meta_data = MetaData(variable_scaling="uv")
>>> pca = create_pca_model(X, meta_data)

here a pca model is created with uv scaling.