Why training time is not decreasing with increase in value of Cost in a SVM model?



I have built one classification model using SVM algorithm.

from sklearn import svm
clf = svm.SVC(kernel='rbf',C=10000)
features_train = features_train[:len(features_train)/100]  # taking only 1% of traing data 
labels_train = labels_train[:len(labels_train)/100] 
print "training time:", round(time()-t0, 3), "s"

I have changed the cost value for calculation of training time at different cost value.

cost - training time
10 - 2.54
100 - 2.56
1000 - 2.45
10000 - 2.42

I want to know why training time is not decreasing with increasing in cost value.


Hi @hinduja1234,

Why do you think that by increasing cost the execution time will reduce?