Consider the table below: Test1 True True False False False True Test2 True True False True False True Test3 False True True True True True Class + + + a. Define the information gain (i.e. reduction of entropy) of a feature F given a set of examples S. Explain all symbols you use. [2 marks] b. Calculate the information gain for each feature in the table above. c. Which test give the best split according to the information gain heuristic? d. Train a Nave Bayesian classifier from the dataset above and predict the class value for the instance X = . Use the m-estimate with uniform prior and m=2 to estimate the probabilities. Show the relevant calculations. [6 marks] [6 marks] [2 marks]