Question

SVM will not perform well with data with more noise because

Answer 1

Answer:

SVM will not perform well with data with more noise because of the weakness of soft margin optimization issue.  The unique hyperplane grabbed in the SVM process using the imbalanced data will be fully skewed towards a minority class. It will actually leads the performance degradation of classifier. With the rise in data imbalance, the overall ration between negative and positive support vectors turn into more imbalanced.