from builtins import range
from ..base import MLClassifierBase
from ..utils import get_matrix_in_format
from sklearn.neighbors import NearestNeighbors
import scipy.sparse as sparse
import numpy as np
class _BinaryRelevanceKNN(MLClassifierBase):
"""Binary Relevance adapted kNN Multi-Label Classifier base class."""
def __init__(self, k=10):
super(_BinaryRelevanceKNN, self).__init__()
self.k = k # Number of neighbours
self.copyable_attrs = ['k']
def fit(self, X, y):
"""Fit classifier with training data
Internally this method uses a sparse CSC representation for y
(:class:`scipy.sparse.csc_matrix`).
Parameters
----------
X : numpy.ndarray or scipy.sparse
input features, can be a dense or sparse matrix of size
:code:`(n_samples, n_features)`
y : numpy.ndaarray or scipy.sparse {0,1}
binary indicator matrix with label assignments.
Returns
-------
self
fitted instance of self
"""
self.train_labelspace = get_matrix_in_format(y, 'csc')
self._n_samples = self.train_labelspace.shape[0]
self._n_labels = self.train_labelspace.shape[1]
self.knn_ = NearestNeighbors(self.k).fit(X)
return self
def predict(self, X):
"""Predict labels for X
Parameters
----------
X : numpy.ndarray or scipy.sparse.csc_matrix
input features of shape :code:`(n_samples, n_features)`
Returns
-------
scipy.sparse of int
binary indicator matrix with label assignments with shape
:code:`(n_samples, n_labels)`
"""
self.neighbors_ = self.knn_.kneighbors(X, self.k, return_distance=False)
self.confidences_ = np.vstack([self.train_labelspace[n, :].tocsc().sum(axis=0) / self.k for n in self.neighbors_])
return self._predict_variant(X)
[docs]class BRkNNaClassifier(_BinaryRelevanceKNN):
"""Binary Relevance multi-label classifier based on k-Nearest Neighbors method.
This version of the classifier assigns the labels that are assigned
to at least half of the neighbors.
Parameters
----------
k : int
number of neighbours
Attributes
----------
knn_ : an instance of sklearn.NearestNeighbors
the nearest neighbors single-label classifier used underneath
neighbors_ : array of arrays of int, shape = (n_samples, k)
k neighbors of each sample
confidences_ : matrix of int, shape = (n_samples, n_labels)
label assignment confidences
References
----------
If you use this method please cite the relevant paper:
.. code :: bibtex
@inproceedings{EleftheriosSpyromitros2008,
author = {Eleftherios Spyromitros, Grigorios Tsoumakas, Ioannis Vlahavas},
booktitle = {Proc. 5th Hellenic Conference on Artificial Intelligence (SETN 2008)},
title = {An Empirical Study of Lazy Multilabel Classification Algorithms},
year = {2008},
location = {Syros, Greece}
}
Examples
--------
Here's a very simple example of using BRkNNaClassifier with a fixed number of neighbors:
.. code :: python
from skmultilearn.adapt import BRkNNaClassifier
classifier = BRkNNaClassifier(k=3)
# train
classifier.fit(X_train, y_train)
# predict
predictions = classifier.predict(X_test)
You can also use :class:`~sklearn.model_selection.GridSearchCV` to find an optimal set of parameters:
.. code :: python
from skmultilearn.adapt import BRkNNaClassifier
from sklearn.model_selection import GridSearchCV
parameters = {'k': range(1,3)}
score = 'f1-macro
clf = GridSearchCV(BRkNNaClassifier(), parameters, scoring=score)
clf.fit(X, y)
"""
def _predict_variant(self, X):
# TODO: find out if moving the sparsity to compute confidences_ boots speed
return sparse.csr_matrix(np.rint(self.confidences_), dtype='i8')
[docs]class BRkNNbClassifier(_BinaryRelevanceKNN):
"""Binary Relevance multi-label classifier based on k-Nearest Neighbors method.
This version of the classifier assigns the most popular m labels of
the neighbors, where m is the average number of labels assigned to
the object's neighbors.
Parameters
----------
k : int
number of neighbours
Attributes
----------
knn_ : an instance of sklearn.NearestNeighbors
the nearest neighbors single-label classifier used underneath
neighbors_ : array of arrays of int, shape = (n_samples, k)
k neighbors of each sample
confidences_ : matrix of int, shape = (n_samples, n_labels)
label assignment confidences
References
----------
If you use this method please cite the relevant paper:
.. code :: bibtex
@inproceedings{EleftheriosSpyromitros2008,
author = {Eleftherios Spyromitros, Grigorios Tsoumakas, Ioannis Vlahavas},
booktitle = {Proc. 5th Hellenic Conference on Artificial Intelligence (SETN 2008)},
title = {An Empirical Study of Lazy Multilabel Classification Algorithms},
year = {2008},
location = {Syros, Greece}
}
Examples
--------
Here's a very simple example of using BRkNNbClassifier with a fixed number of neighbors:
.. code :: python
from skmultilearn.adapt import BRkNNbClassifier
classifier = BRkNNbClassifier(k=3)
# train
classifier.fit(X_train, y_train)
# predict
predictions = classifier.predict(X_test)
You can also use :class:`~sklearn.model_selection.GridSearchCV` to find an optimal set of parameters:
.. code :: python
from skmultilearn.adapt import BRkNNbClassifier
from sklearn.model_selection import GridSearchCV
parameters = {'k': range(1,3)}
score = 'f1-macro
clf = GridSearchCV(BRkNNbClassifier(), parameters, scoring=score)
clf.fit(X, y)
"""
def _predict_variant(self, X):
avg_labels = [int(np.average(self.train_labelspace[n, :].sum(axis=1)).round()) for n in self.neighbors_]
prediction = sparse.lil_matrix((X.shape[0], self._n_labels), dtype='i8')
top_labels = np.argpartition(self.confidences_, kth=min(avg_labels + [len(self.confidences_[0])]),
axis=1).tolist()
for i in range(X.shape[0]):
for j in top_labels[i][-avg_labels[i]:]:
prediction[i, j] += 1
return prediction