Given a general problem $(i=1,2,$dotsN$)$ of $m$ dimensional features labeled as $L$

number of classes, it was quested to design an SVM solution for classification

task.

a figure is required to show how to design the classifier:

How you plan to use the data given? Such training, validation and test set?

What are the key parameters in SVM you need to tune?

What is your plan to achieve the optimized solution?

How are you going to test the robustness of your model?