Exercise 1: Supervised classification 1)Why is pruning useful for induction from decision trees? 2)What are the 2 common approaches to decision tree pruning? Explain their principle. 3)Compare the advantages and disadvantages of "eager" type classification (e.g. decision tree, rules, Bayesian network) versus "lazy" type classification (e.g. k-nearest neighbor, case-based reasoning). 4)Given the following relationship "weather.symbolic" where the class attribute is "Play". Which attributes are relevant for classification? Explain why. 5)Considering that the learning set contains the first 8 instances {O1 to O8}. Use the naive Bayesian classification technique with Laplace's formula and relevant attributes to predict the class of each of the last 6 instances {O9 toO14}. 6)Construct the confusion matrix. 7)Determine the error rate, accuracy and recall. 8)What type of evaluation method is used here?

@relation weather.symbolic @attribute name nominal @attribute outlook { sunny, overcast, rainy} @attribute temperature {hot, mild, cool} @attribute humidity {high, normal} @attribute windy { TRUE, FALSE} @attribute play {yes, no} @data 01, overcast, hot, high, FALSE, yes 02, rainy, mild, high, FALSE, yes 03, rainy, cool, normal, FALSE, yes 04, overcast, cool, normal, TRUE, yes 05, sunny, cool, normal, FALSE, 06, sunny, hot, high, FALSE, no 07, sunny, high, TRUE, no 08, rainy, cool, normal, TRUE, no 09, rainy, mild, normal, FALSE, yes 010, sunny, mild, normal, TRUE, yes 011, sunny, mild, high, FALSE, no 012, overcast, mild, high, TRUE, yes 013, overcast, hot, normal, FALSE, 014, rainy, mild, high, TRUE, no yes hot, yes @relation weather.symbolic @attribute name nominal @attribute outlook { sunny, overcast, rainy} @attribute temperature {hot, mild, cool} @attribute humidity {high, normal} @attribute windy { TRUE, FALSE} @attribute play {yes, no} @data 01, overcast, hot, high, FALSE, yes 02, rainy, mild, high, FALSE, yes 03, rainy, cool, normal, FALSE, yes 04, overcast, cool, normal, TRUE, yes 05, sunny, cool, normal, FALSE, 06, sunny, hot, high, FALSE, no 07, sunny, high, TRUE, no 08, rainy, cool, normal, TRUE, no 09, rainy, mild, normal, FALSE, yes 010, sunny, mild, normal, TRUE, yes 011, sunny, mild, high, FALSE, no 012, overcast, mild, high, TRUE, yes 013, overcast, hot, normal, FALSE, 014, rainy, mild, high, TRUE, no yes hot, yes