$(4$ points$)$ This is a real dataset from a credit union $($data here darr, in which the charge$-$off information for $30,794$ customers are kept together with three yariables, total monthly income, credit score, current years employed. Charge$-$off $($or bad accounts$)$ is defined as those customers who defaulted on $($failed to pay for$)$ their han from shank Good hecounts are those customers who are current in their payment. Apply the same intuitive approach introduced in the cancer prediction dataset on this larger dataset to answer the same questions:

$(10)-$ How eventh tell if any of these three variable can predict the risk of charge$-$off?

If so$,$ can we tell which one is most predictive?

If so$,$ how can we produce an estimate of risk of charge$-$off based on these variables? you choose range properly to best show the distributions. Use number of groups $=40$ for Intuition $2.$