You suspect the SVM is underfitting your dataset. Which of the following methods should you adopt to ensure better results?

#1

Increasing the value of C

Decreasing the value of C

Keeping the value of C the same but shuffling the data

Re-initialising the initial weights