Multiclass 分类

Multiclass classification is a classification task with more than two classes. Each sample can only be labeled as one class.

目标变量格式

• 包含多个离散值的一维向量

import numpy as np

y = np.array(["apple", "pear", "apple", "orange"])

• 形状为 $(n\_samples, n\_classes)$ 的二元矩阵

import numpy as np
from sklearn.preprocessing import LabelBinarizer
from scipy import sparse

y = np.array(["apple", "pear", "apple", "orange"])

# dense binary matrix
y_dense = LabelBinarizer().fit_transform(y)
y_dense

# sparse binary matrix
y_sparse = sparse.csr_matrix(y_dense)
y_sparse

 [[1 0 0]
  [0 0 1]
  [1 0 0]
  [0 1 0]]

(0, 0)        1
(1, 2)        1
(2, 0)        1
(3, 1)        1

OneVsRestClassifier

one-vs-rest，one-vs-all

from sklearn import datasets
from sklearn.multiclass import OneVsRestClassifier
from sklearn.svm import LinearSVC

X, y = datasets.load_iris(return_X_y = True)
svc = LinearSVC(random_state = 0)
OVR = OneVsRestClassifier(svc).fit(X, y).predict(X)

OneVsOneClassifier

from sklearn import datasets
from sklearn.multiclass import OneVsOneClassifier
from sklearn.svm import LinearSVC

X, y = datasets.load_iris(return_X_y = True)
svc = LinearSVC(random_state = 0)
OVO = OneVsOneClassifier(svc).fit(X, y).predict(X)

OutputCodeClassifier

Multilabel 分类

Multilabel classification

Multiclass-Multioutput 分类

Multiclass-Multioutput classification

Multioutput 回归

Multioutput regression

目标变量格式

import numpy as np

y = np.array(
    [[31.4, 94],
     [40.5, 109],
     [25.0, 30]]
)

MultiOutputRegressor

from sklearn.datasets import make_regression
from sklearn.multioutput import MultiOutputRegressor
from sklearn.ensemble import GradientBoostingRegressor

X, y = make_regression(n_samples = 10, n_targets = 3, random_state = 1)

gbr = GradientBoostingRegressor(random_state = 0)
MOR = MultiOutputRegressor(gbr).fit(X, y).predict(X)

RegressorChain

参考

• Multiclass and multioutput algorithms