b) We would like to predict the sex of a person based on two binary atuributes: leg-cover (pants or skirts) and facial-hair (some or none). We have a data set of 2,000 individuals, half male and half female. 75% of the males have no facial hair. Skirts are wom by 50%/ of the females. All females are barefaced and no male wears a skirt. For your information, log2(1/8)m3.00,log2(1/4)=2,00,log2(1/3)=1.58,log(3/8)= 1.42,log2(3/7)=1.22,log2(1/2)=1.00,log2(4/T)=0.811log3(5/8)=0.68, log2(2/3)=0.58,log2(3/4)=0.42,log2(7/8)=0.19 (i) What is the iritial entropy in the system? (2 maris) (ii) Compute the information gain of initially choosing the attribute leg-cover and for initially choosing facial-hair. (4 marks) (iii) Based on yours answers, which attribute should be used us the root of a decision tree? (2 marks)