ITMO_FS.filters.unsupervised.TraceRatioLaplacian

class ITMO_FS.filters.unsupervised.TraceRatioLaplacian(n_selected_features, k=5, t=1)

Creates TraceRatio(similarity based) feature selection filter performed in unsupervised way, i.e laplacian version

Parameters:
  • n_selected_features (int) – Amount of features to filter
  • k (int) – number of neighbours to use for knn
  • t (int) –

    constant for kernel function calculation

    • Note: in laplacian case only. In fisher it uses label similarity, i.e if both samples belong to same class

Notes

For more details see this paper.

Examples

>>> from ITMO_FS.filters.unsupervised.trace_ratio_laplacian import TraceRatioLaplacian
>>> from sklearn.datasets import make_classification
>>> x, y = make_classification(1000, 100, n_informative = 10, n_redundant = 30, n_repeated = 10, shuffle = False)
>>> tracer = TraceRatioLaplacian(10)
>>> print(tracer.run(x, y)[0])
__init__(n_selected_features, k=5, t=1)

Initialize self. See help(type(self)) for accurate signature.

run(X, y)

Fits filter

Parameters:
  • X (numpy array, shape (n_samples, n_features)) – The training input samples
  • y (numpy array, shape (n_samples, )) – The target values
Returns:

feature_indices – array of feature indices in X
Return type:

numpy array

Examples