You are trying to create a model that will predict which companies will be good or bad investments. The outcomes are negative (bad investment) or positive (good investment). Currently you are using two features: price (low, high) and industry E {food, tech). Here is the training data: price industry outcome high tech X low food as high tech 24 low tech tech food food as x6 low] low high (a) Create a single feature decision tree (decision stump) for the feature price. At the leaves, clearly label the probability of a positive outcome (Ppos). (b) Create a single feature decision tree (decision stump) for the feature industry. At the leaves, clearly label the probability of a positive outcome (Ppos). (c) Using the threshold 0.2, which leaves would result in a positive classification result for a test example? (d) Using the threshold 0.6, which leaves would result in a positive classification result for a test example? (Separate questions from the previous page) (e) For the single discrete (categorical) feature below (type of fruit), how could you convert this feature into numerical feature(s)? Transform the data (here n = 6) using the right side of the grid below, and briefly explain your answer. fruit I1 orange I2 pear I3 pear apple 4 IS I6 orange apple (f) Using big-O notation, how long does it take to create a single feature decision tree model (decision stump)? You may assume that: n is the number of training examples, m is the number of testing examples, all example have p features, the feature in question has u possible values, and the outcome is binary. Explain your reasoning (including any assumptions about implementation) for full credit. (g) Considering the same assumptions as above, how long does it take to predict the outcomes for the m testing examples? Again answer using big-O notation and explain your reasoning.