cliu78 / classifier_machinelearning Goto Github PK

Classifier_MachineLearning -- use JAVA and the weka API General Classification Use the MultilayerPerceptron classifier (use the default configuration that is set by the no argument constructor) and the dataset glass.arff. a) Train the classifier on the first N percent of instances. The testing data will be the last 100-N percent of the instances. i) For each N={10,30,50,70,90} report the accuracy of the classifier over the training data and the accuracy over the testing data. ii) Plot accuracy on training data vs. N and accuracy on testing data vs. N on the same plot b) Use n-fold cross validation to evaluate the accuracy (as a percentage) of the classifier over the dataset. Do this for n={2,5,10,15,20}. Hint: The Evaluation class in weka can do cross validation. i) Plot the accuracy vs n. ii) Plot the time taken by cross validation vs. n. see more in readme.