0
$\begingroup$

I have multilabel problem.
I was using onevsrestclassifier and now i want to use onevsoneclassifier.

from sklearn.svm import SVC
classifire = SVC(gamma='auto')
accuracy = []

skf = StratifiedKFold(n_splits=10,random_state=None)
kf = KFold(n_splits=5)

skf.get_n_splits(X_train,y_train)

for train_index,test_index in kf.split(X_train,y_train):
    X1_train,X1_test = X_train.iloc[train_index],X_train.iloc[test_index]
    y1_train,y1_test = y_train.iloc[train_index],y_train.iloc[test_index]
    clf = OneVsOneClassifier(classifire)
#     clf = OneVsRestClassifier(classifire)
    clf.fit(X1_train,y1_train)
    prediction = clf.predict(X1_test)
    
    score = accuracy_score(prediction,y1_test)
    print(score)

But i am getting error only in OneVsOneClassifier

<ipython-input-47-91b351c58c24> in <module>
     11     clf = OneVsOneClassifier(classifire)
     12 #     clf = OneVsRestClassifier(classifire)
---> 13     clf.fit(X1_train,y1_train)
     14     prediction = clf.predict(X1_test)
     15 
..
ValueError: bad input shape (19051, 206)

I am new here. Can you help me ?

print(X1_train.shape)
print(y1_train.shape)
print(X1_test.shape)
print(y1_test.shape)

o/p

(19051, 875)
(19051, 206)
(4763, 206)
(4763, 875)

UPDATE

accuracy = []

skf = StratifiedKFold(n_splits=10,random_state=None)
kf = KFold(n_splits=5)

skf.get_n_splits(X_train,y_train)

for train_index,test_index in kf.split(X_train,y_train):
    X1_train,X1_test = X_train.iloc[train_index],X_train.iloc[test_index]
    y1_train,y1_test = y_train.iloc[train_index],y_train.iloc[test_index]
    clf = OneVsOneClassifier(classifire)
#     clf = OneVsRestClassifier(classifire)
    print("X1_train",X1_train.shape)
    print("y1_train",y1_train.shape)
    print('X1_test',X1_test.shape)
    print('y1_test',y1_test.shape)
    clf.fit(X1_train,y1_train)
    prediction = clf.predict(X1_test)
    
    score = accuracy_score(prediction,y1_test)
    print(score)

And the output is

X1_train (19051, 875)
y1_train (19051, 206)
X1_test (4763, 875)
y1_test (4763, 206)

So X1_train and test have same columns. But the error is same

Improve this question
$\endgroup$
7
  • $\begingroup$ the rows of X1_train and y1_train should match, before the error line type this command and print the output - print(X1_train.shape,y1_train.shape) $\endgroup$
    – Madhur Yadav
    Sep 21, 2020 at 12:00
  • $\begingroup$ also print -> print(X_train.shape,y_train.shape) $\endgroup$
    – Madhur Yadav
    Sep 21, 2020 at 12:02
  • 1
    $\begingroup$ X train and test should have same number of columns and (X train and Y train) should have same number of rows and (Xtest and Ytest) should have same number of rows $\endgroup$
    – Madhur Yadav
    Sep 21, 2020 at 14:24
  • 1
    $\begingroup$ your x_train and X_test doesn't have same number of columns(input features) $\endgroup$
    – Madhur Yadav
    Sep 21, 2020 at 14:25
  • 1
    $\begingroup$ Vohoooo thankyou man. $\endgroup$
    – Rajan Lagah
    Sep 21, 2020 at 14:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.