a$)$ The data science team constructed two models by using the same training data, namely

Model A and Model B$.$ Both of these models are illustrated below.

Model A uses attributes BMI, Temperature and Age to form the tree. In the leaf nodes, the

number of patients belonging to each class is given. For instance, when BMI category equals to

High and Temperature category is Normal, in the leaf node, $100$ patients had a cancer $($category

yes$)$ and $150$ patients did not have a cancer $($category no$).$

Model B uses attributes BMI, Smoke and Temperature to form the tree. Similarly, in the leaf

nodes, the number of patients belonging to each target variable category is given. For instance,

when BMI category equals to High and Smoke category equals to No$,$ in the leaf node, $60$ patients

had a cancer $($category yes$)$ and $140$ patients did not have a cancer $($category no$).$

Calculate the frequency based estimation probabilities $($or target variable class prediction

probabilities$)$ in both of these decision tree models. Use these probabilities as the class

prediction probabilities to predict the class in each leaf $($or terminal$)$ node with the threshold

value $0\mathrm{.}5.$ Which model should be used for predicting the pancreatic cancer? Explain your advice.