Consider a classification problem with general K classes. Assume that the data set is standardized within each class. The goal is to use the naive Bayes classification method with beta distribution as the conditional distribution of the inputs given the class. You must use the maximum likelihood estimators of the within class parameters. 1) Find log P(G = k X) P (G = KIX) for k = 1,2, .., K - 1. 2) expand the answer of (1) to be in the form of Generalized Additive Model (GAM) and identify the constants ax, and find the functions gk.i. 3) find the discriminant function, 4) find the decision boundary of the classification problem, 5) write down your own function in R of the naive Bayes classification using the beta distribution for general K. That function shouldn't use any special package for naive Bayes classification. 6) Apply that function to the iris data in problem 2. Is there any improvement in the classification when compared to results of Problem 2