tick.preprocessing.
LongitudinalFeaturesLagger
(n_lags, n_jobs=-1)[source]¶Transforms longitudinal exposure features to add columns representing lagged features.
This preprocessor transform an input list of n_cases
numpy ndarrays or
scipy.sparse.csr_matrices of shape (n_intervals, n_features)
so as to
add columns representing the lagged exposures. It outputs a list of
n_cases
numpy arrays or csr_matrices of shape
(n_intervals, n_features + sum(n_lags + 1))
.
Exposure can take two forms: - short repeated exposures: in that case, each column of the numpy arrays or csr matrices can contain multiple ones, each one representing an exposure for a particular time bucket. - infinite unique exposures: in that case, each column of the numpy arrays or csr matrices can only contain a single one, corresponding to the starting date of the exposure.
n_lags : numpy.ndarray
, shape=(n_features,), dtype=”uint64”
Number of lags to compute: the preprocessor adds columns representing lag = 1, …, n_lags[i] for each feature [i]. If
n_lags
is a null vector, this preprocessor does nothing.n_lags
must be non-negative.
Examples
>>> import numpy as np
>>> from scipy.sparse import csr_matrix
>>> from tick.preprocessing.longitudinal_features_lagger import LongitudinalFeaturesLagger
>>> features = [csr_matrix([[0, 1, 0],
... [0, 0, 0],
... [0, 0, 1]], dtype="float64"),
... csr_matrix([[1, 1, 0],
... [0, 0, 1],
... [0, 0, 0]], dtype="float64")
... ]
>>> censoring = np.array([3, 2], dtype="uint64")
>>> n_lags = np.array([2, 1, 0], dtype='uint64')
>>> lfl = LongitudinalFeaturesLagger(n_lags)
>>> product_features, _, _ = lfl.fit_transform(features)
>>> # output comes as a list of sparse matrices or 2D numpy arrays
>>> product_features.__class__
<class 'list'>
>>> [x.toarray() for x in product_features]
[array([[0., 0., 0., 1., 0., 0.],
[0., 0., 0., 0., 1., 0.],
[0., 0., 0., 0., 0., 1.]]), array([[1., 0., 0., 1., 0., 0.],
[0., 1., 0., 0., 1., 1.],
[0., 0., 1., 0., 0., 0.]])]
fit
(features, labels=None, censoring=None)[source]¶Fit the feature lagger using the features matrices list.
features : list
of numpy.ndarray
or list
of scipy.sparse.csr_matrix
,
list of length n_cases, each element of the list of shape=(n_intervals, n_features) The list of features matrices.
censoring : numpy.ndarray
, shape=(n_cases,), dtype=”uint64”
The censoring data. This array should contain integers in [1, n_intervals]. If the value i is equal to n_intervals, then there is no censoring for sample i. If censoring = c < n_intervals, then the observation of sample i is stopped at interval c, that is, the row c - 1 of the corresponding matrix. The last n_intervals - c rows are then set to 0.
output : LongitudinalFeaturesLagger
The fitted current instance.
transform
(features, labels=None, censoring=None)[source]¶Add lagged features to the given features matrices list.
features : list
of numpy.ndarray
or list
of scipy.sparse.csr_matrix
,
list of length n_cases, each element of the list of shape=(n_intervals, n_features) The list of features matrices.
censoring : numpy.ndarray
, shape=(n_cases,), dtype=’uint64’, default=’None’
The censoring data. This array should contain integers in [1, n_intervals]. If the value i is equal to n_intervals, then there is no censoring for sample i. If censoring = c < n_intervals, then the observation of sample i is stopped at interval c, that is, the row c - 1 of the correponding matrix. The last n_intervals - c rows are then set to 0.
output : [numpy.ndarrays]
or [csr_matrices]
, shape=(n_intervals, n_features)
The list of features matrices with added lagged features.