使用Scikit-learn训练SVM（支持向量机）分类器

Question

I want to train different classifier with using Scikit-learn with following code for Multi-label classification problem:

names = [
    "Nearest Neighbors",
    "Linear SVM", "RBF SVM", "Gaussian Process",
    "Decision Tree", "Random Forest", "Neural Net", "AdaBoost",
    "Naive Bayes", "QDA"]

classifiers = [
    KNeighborsClassifier(3),
    SVC(C=0.025),
    SVC(gamma=2, C=1),
    GaussianProcessClassifier(1.0 * RBF(1.0)),
    DecisionTreeClassifier(max_depth=5),
    RandomForestClassifier(max_depth=5),
    MLPClassifier(alpha=0.5),
    AdaBoostClassifier(),
    GaussianNB(),
    QuadraticDiscriminantAnalysis()]

for name, clf in izip(names, classifiers):
    clf.fit(X_train, Y_train)
    score = clf.score(X_train, Y_test)
    print name, score

KNeighbors classifier works properly but when I reach to the SVM classifier it throws following exception:

Traceback (most recent call last):
  File "/Users/mac/PycharmProjects/GraphLstm/classifier.py", line 87, in <module>
    clf.fit(X_train, Y_train)
  File "/Library/Python/2.7/site-packages/sklearn/svm/base.py", line 151, in fit
    X, y = check_X_y(X, y, dtype=np.float64, order='C', accept_sparse='csr')
  File "/Library/Python/2.7/site-packages/sklearn/utils/validation.py", line 526, in check_X_y
    y = column_or_1d(y, warn=True)
  File "/Library/Python/2.7/site-packages/sklearn/utils/validation.py", line 562, in column_or_1d
    raise ValueError("bad input shape {0}".format(shape))
ValueError: bad input shape (9280, 39)

What's the reason and How can I fix that?

Edit: As commented by @Vivek following classifier only allowed for Multi-label classification :

sklearn.tree.DecisionTreeClassifier
sklearn.tree.ExtraTreeClassifier
sklearn.ensemble.ExtraTreesClassifier
sklearn.neighbors.KNeighborsClassifier
sklearn.neural_network.MLPClassifier
sklearn.neighbors.RadiusNeighborsClassifier
sklearn.ensemble.RandomForestClassifier
sklearn.linear_model.RidgeClassifierCV

Answer 1

The fit function of the knn classifier allows a matrix as y-value. For the svm this is not allowed. The error message tries to hint you on a disallowed y-shape

Answer 2

Since this is a multi-label classification problem, not all estimators in scikit will be able to handle them inherently. The documentation provides a list of estimators which support multi-label out of the box like various tree based estimators or others :

sklearn.tree.DecisionTreeClassifier
sklearn.tree.ExtraTreeClassifier
sklearn.ensemble.ExtraTreesClassifier
sklearn.neighbors.KNeighborsClassifier
...
...

However there are strategies like one-vs-all which can be employed to train the required estimator (which doesn't support multilabel directly). Sklearn estimator OneVsRestClassifier is made for this.

See the documentation here for more details about it.