You will build classification binary output models using all five of the following algorithms/techniques using R program. In each of the analyses, use the same dependent variable and the same independent variables (although depending on the algorithm, some independent variables may require different preprocessing). You will compare their predictive ability and note the strengths and weaknesses of each approach.

? Logistic regression ? k-nearest neighbors (and find an optimal value for k) ? Nave Bayes ? Decision tree (run your model with different complexity parameters that result in variations on the decision tree) ? Neural network (use one hidden layer with 3 nodes) (remember that all independent variables used in your neural network should be equally scaled)

Dataset:

A B D G K Marital st: Daytime/ Previous ( Mother's ( Father's q Education Debtor Tuition fe Gender Age at ent Target 1 Dropout W N O 1 Dropout 1 Dropout 1 Dropout - HP 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout + + $ U U U G NN NNNNNNNN 1 Dropout 1 Dropout 3 + + $ A U N N N O W P U P W W W W W W W N N 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout -OOO 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 1 Dropout 33 1 Dropout 34 1 Dropout 35 1 Dropout