$-$Differentiate between LinearSVC, SVC$,$ and SGDClassifier in the scikit$-$learn library. Highlight the key distinctions among these classes and explain how to select the most suitable class for a machine learning challenge.

$-$If you have trained an SVM classifier using an RBF kernel, and it appears to underfit the training data, can you explain how you would adjust the parameters to improve the model? Would you increase or decrease the gamma parameter, and how would changes to the C parameter affect the model$$s performance? Provide reason focusing on the impact of each parameter on the model$$s complexity and its ability to fit the training set more accurately.