The goal behind analyzing this dataset is to determine which parameters yield the highest prediction results. Sources such as GoogleAdWord, GoogleSearch and Directlink are observed to see out of these three generate the highest revenue. The mean tells us that using the linear model for GoogleAdWords gave a better prediction rate of 0.500. GoogleSearch had a prediction rate of 3.07 and Directlink gave a prediction rate of 155.396. The methods use for this predictin model inovlved permutation testing, decision tree, cross-validation and z-test to determine the outcome for the above mentioned.
Machine Learning packages used: Rpart, SVM, MSE, LM
Other Methods used: Permutation testing, Decision tree, Cross-validation, Z-test