or this assignment, you'll be working with the WEKA datamining software: http://www.cs.waikato.ac.nz/ml/weka/ What to turn in: Turn in your answers to the questions below. ALSO turn in the text output from WEKA for each of your experiments (at the bottom of the same file). 1. Work through this initial classification tutorial for Weka: http://machinelearningmastery.com/how-to-run-your-first-classifier-in-weka/ Why was the confusion matrix 3x3 instead of 2x2? 2. Now try using the Lazy "IBk" algorithm (K-Nearest Neighbor) with the Iris dataset with 10-fold cross validation. Use several values of k and attempt to find an optimal accuracy. What was the best value of k that you found, and what was its accuracy? 3. Now try the Iris dataset with the Naive Bayes classifier. Briefly compare the results with previous techniques. Also, take a look through the description of this plugin and its options and explain how it deals with the zero probability problem. 4. Now try the MultilayerPerceptron to apply a feed-forward neural network model to the Iris data set. Try different network architectures (numbers of hidden layers, nodes per layer). What worked the best for you, and why? page 1