Question 9 (8 marks) Consider the selection of training examples shown in the following table. This data corresponds to nine decisions (Action) that were made in different situations. The decisions recorded here reflects the experts judgment of whether to release or hold the deployment of a specialised tracking balloon that will be able to monitor and record certain atmospheric weather conditions. Observation Overcast StrongWinds UV_ratio Action 1 yes yes 1.0 release 2 yes yes 6.0 release 3 yes no 5.0 hold 4 no no 4.0 release 5 no yes 7.0 hold 6 yes 3.0 hold 7 no 8.0 hold 8 yes no 7.0 release 9 no yes hold no no 5.0 (a) What is the entropy from this data with respect to the decision for releasing balloons? (2 marks) (b) What are the information gains for the conditions of "Overcast" and "Strong Winds" relative to this selection of training examples? (3 marks) (c) For the continuous attribute, "UV_ratio" compute the information gain for every possible split, and thus determine which attribute from Overcast, Strong Winds and UV_ratio which provides the best split according to the information gain