NMF

class ibex.sklearn.decomposition.NMF(n_components=None, init=None, solver='cd', beta_loss='frobenius', tol=0.0001, max_iter=200, random_state=None, alpha=0.0, l1_ratio=0.0, verbose=0, shuffle=False)

Bases: sklearn.decomposition.nmf.NMF, ibex._base.FrameMixin

Note

The documentation following is of the class wrapped by this class. There are some changes, in particular:

Note

The documentation following is of the original class wrapped by this class. This class wraps the attribute components_.

Example:

>>> import pandas as pd
>>> import numpy as np
>>> from ibex.sklearn import datasets
>>> from ibex.sklearn.decomposition import PCA as PdPCA

>>> iris = datasets.load_iris()
>>> features = iris['feature_names']
>>> iris = pd.DataFrame(
...     np.c_[iris['data'], iris['target']],
...     columns=features+['class'])

>>> iris[features]
sepal length (cm)  sepal width (cm)  petal length (cm)  petal width (cm)
0                5.1               3.5                1.4               0.2
1                4.9               3.0                1.4               0.2
2                4.7               3.2                1.3               0.2
3                4.6               3.1                1.5               0.2
4                5.0               3.6                1.4               0.2
...

>>> PdPCA(n_components=2).fit(iris[features], iris['class']).transform(iris[features])
    comp_0    comp_1
0   -2.684207 ...0.326607
1   -2.715391 ...0.169557
2   -2.889820 ...0.137346
3   -2.746437 ...0.311124
4   -2.728593 ...0.333925
...

Non-Negative Matrix Factorization (NMF)

Find two non-negative matrices (W, H) whose product approximates the non- negative matrix X. This factorization can be used for example for dimensionality reduction, source separation or topic extraction.

The objective function is:

0.5 * ||X - WH||_Fro^2
+ alpha * l1_ratio * ||vec(W)||_1
+ alpha * l1_ratio * ||vec(H)||_1
+ 0.5 * alpha * (1 - l1_ratio) * ||W||_Fro^2
+ 0.5 * alpha * (1 - l1_ratio) * ||H||_Fro^2

Where:

||A||_Fro^2 = \sum_{i,j} A_{ij}^2 (Frobenius norm)
||vec(A)||_1 = \sum_{i,j} abs(A_{ij}) (Elementwise L1 norm)

For multiplicative-update (‘mu’) solver, the Frobenius norm (0.5 * ||X - WH||_Fro^2) can be changed into another beta-divergence loss, by changing the beta_loss parameter.

The objective function is minimized with an alternating minimization of W and H.

Read more in the User Guide.

n_components : int or None
Number of components, if n_components is not set all features are kept.
init : ‘random’ | ‘nndsvd’ | ‘nndsvda’ | ‘nndsvdar’ | ‘custom’

Method used to initialize the procedure. Default: ‘nndsvd’ if n_components < n_features, otherwise random. Valid options:

  • ‘random’: non-negative random matrices, scaled with:
    sqrt(X.mean() / n_components)
  • ‘nndsvd’: Nonnegative Double Singular Value Decomposition (NNDSVD)
    initialization (better for sparseness)
  • ‘nndsvda’: NNDSVD with zeros filled with the average of X
    (better when sparsity is not desired)
  • ‘nndsvdar’: NNDSVD with zeros filled with small random values
    (generally faster, less accurate alternative to NNDSVDa for when sparsity is not desired)
  • ‘custom’: use custom matrices W and H
solver : ‘cd’ | ‘mu’

Numerical solver to use: ‘cd’ is a Coordinate Descent solver. ‘mu’ is a Multiplicative Update solver.

New in version 0.17: Coordinate Descent solver.

New in version 0.19: Multiplicative Update solver.

beta_loss : float or string, default ‘frobenius’

String must be in {‘frobenius’, ‘kullback-leibler’, ‘itakura-saito’}. Beta divergence to be minimized, measuring the distance between X and the dot product WH. Note that values different from ‘frobenius’ (or 2) and ‘kullback-leibler’ (or 1) lead to significantly slower fits. Note that for beta_loss <= 0 (or ‘itakura-saito’), the input matrix X cannot contain zeros. Used only in ‘mu’ solver.

New in version 0.19.

tol : float, default: 1e-4
Tolerance of the stopping condition.
max_iter : integer, default: 200
Maximum number of iterations before timing out.
random_state : int, RandomState instance or None, optional, default: None
If int, random_state is the seed used by the random number generator; If RandomState instance, random_state is the random number generator; If None, the random number generator is the RandomState instance used by np.random.
alpha : double, default: 0.

Constant that multiplies the regularization terms. Set it to zero to have no regularization.

New in version 0.17: alpha used in the Coordinate Descent solver.

l1_ratio : double, default: 0.

The regularization mixing parameter, with 0 <= l1_ratio <= 1. For l1_ratio = 0 the penalty is an elementwise L2 penalty (aka Frobenius Norm). For l1_ratio = 1 it is an elementwise L1 penalty. For 0 < l1_ratio < 1, the penalty is a combination of L1 and L2.

New in version 0.17: Regularization parameter l1_ratio used in the Coordinate Descent solver.

verbose : bool, default=False
Whether to be verbose.
shuffle : boolean, default: False

If true, randomize the order of coordinates in the CD solver.

New in version 0.17: shuffle parameter used in the Coordinate Descent solver.

components_ : array, [n_components, n_features]
Factorization matrix, sometimes called ‘dictionary’.
reconstruction_err_ : number
Frobenius norm of the matrix difference, or beta-divergence, between the training data X and the reconstructed data WH from the fitted model.
n_iter_ : int
Actual number of iterations.
>>> import numpy as np
>>> X = np.array([[1, 1], [2, 1], [3, 1.2], [4, 1], [5, 0.8], [6, 1]])
>>> from sklearn.decomposition import NMF
>>> model = NMF(n_components=2, init='random', random_state=0)
>>> W = model.fit_transform(X)
>>> H = model.components_

Cichocki, Andrzej, and P. H. A. N. Anh-Huy. “Fast local algorithms for large scale nonnegative matrix and tensor factorizations.” IEICE transactions on fundamentals of electronics, communications and computer sciences 92.3: 708-721, 2009.

Fevotte, C., & Idier, J. (2011). Algorithms for nonnegative matrix factorization with the beta-divergence. Neural Computation, 23(9).

fit(X, y=None, **params)[source]

Note

The documentation following is of the class wrapped by this class. There are some changes, in particular:

Learn a NMF model for the data X.

X : {array-like, sparse matrix}, shape (n_samples, n_features)
Data matrix to be decomposed

y : Ignored.

self

fit_transform(X, y=None, W=None, H=None)[source]

Note

The documentation following is of the class wrapped by this class. There are some changes, in particular:

Learn a NMF model for the data X and returns the transformed data.

This is more efficient than calling fit followed by transform.

X : {array-like, sparse matrix}, shape (n_samples, n_features)
Data matrix to be decomposed

y : Ignored.

W : array-like, shape (n_samples, n_components)
If init=’custom’, it is used as initial guess for the solution.
H : array-like, shape (n_components, n_features)
If init=’custom’, it is used as initial guess for the solution.
W : array, shape (n_samples, n_components)
Transformed data.
inverse_transform(W)[source]

Note

The documentation following is of the class wrapped by this class. There are some changes, in particular:

Transform data back to its original space.

W : {array-like, sparse matrix}, shape (n_samples, n_components)
Transformed data matrix
X : {array-like, sparse matrix}, shape (n_samples, n_features)
Data matrix of original shape

New in version 0.18.

transform(X)[source]

Note

The documentation following is of the class wrapped by this class. There are some changes, in particular:

Transform the data X according to the fitted NMF model

X : {array-like, sparse matrix}, shape (n_samples, n_features)
Data matrix to be transformed by the model
W : array, shape (n_samples, n_components)
Transformed data