A mobile network operator company is training a machine learning model to predict. customer churn using Stochastic Gradient Descent $($SGD$).$ The training dataset consists of$.$ of $10,000$ samples. $($a$)$ Describe the initial setup for using SGD$,$ including the significance using the mini$-$batch size and learning rate? $($b$)$ Identify one potential challenge when using SGD and suggest a solution to address it$?($c$)$ State three variants of Stochastic Gradient Descent?