RStudio lang.

Jan 28th, 2018 (11:59 PM) instruction: - Late submissions are not allowed. - Submit working R code as a “ .R” file or submit an html notebook generated using R. 1. The dataset used in this problem is related to the quality of white wine. The data includes 4898 samples of wine. There are 11 variables have potential effect to the response variable (quality). The data file “[login to view URL]” and its description “[login to view URL]” are uploaded on Canvas. The goal of this problem is to get you familiar with R and using R-packages to make the decision tree model and random forest model. a) Using “rpart” R-package to make a decision tree and plot the tree. Which variable has the highest impact to the quality of white wine (show your result)? b) Making a random forest model using the whole dataset, report your results. How does this result look like in comparison to part a? 2. The file “[login to view URL]” is a dataset about how effective the ads on social media are. The data has three variables (Gender, Age, and Estimated Salary) and one response variable (Purchased). This response variable has two value 0 (no purchased) and 1 (purchased successfully) (Let’s consider 0 and 1 as factor, not number). a) Split the data into training (75% of the data) and testing set (25% of data left) b) Build a decision tree model for the training set. Predict the response value for the testing set. Report your confusion matrix for the prediction and plot the decision tree. c) Build a random forest model for the training set. Predict the response value for the testing set. Report your confusion matrix for the prediction. Compare the result with part b? Is there any difference? Explain? Data sources: 1. http://archive.ics.uci.edu/ml/datasets/Wine+Quality 2. [login to view URL]