CODEWITHSUNDEEP
pythonscikit-learnsvm
Xudong
Xudong

Reputation: 525

Why LinearSVC works so bad on this dataset?

I just generated a dataset, and from the scatter plot I think these scatters are quite separable. But I really don't know why the Linear SVC-sklearn worked very bad. The dataset scatters plot and the SVC result are below:

scatter plot Linear SVC result

The code I use is like

from sklearn.svm import LinearSVC    
svc = LinearSVC()    
model = svc.fit(X, y)

Could anyone help me with that problem? I really think SVM should have a better result.

The dataset csv file is here: data csv file

The first two columns are x and y respectively, and the third column is the data label.

Upvotes: 0

Views: 370

Answers (1)

Juan Carlos Ramirez
Juan Carlos Ramirez

Reputation: 2129

Assuming your data is centered set your fit_intercept to False. You might do better here with less regularization, by increasing the cost of the error (C=100): 

svc = LinearSVC(fit_intercept=False, C=100)

Upvotes: 2

Related Questions