HW2 - Accurate solution PDF

Preview

Summary

Homework 2: System Administrator PerformanceWhat to Turn In:1) An MS Word file that contains your answers to the questions. You can use screenshots toreport the required results if it is convenient for you.2) An R script that contains your code.Problem: Predicting System Administrator PerformanceA m...

Description

Homework 2: System Administrator Performance What to Turn In: 1) An MS Word file that contains your answers to the questions. You can use screenshots to report the required results if it is convenient for you. 2) An R script that contains your code.

Problem: Predicting System Administrator Performance A management consultant is studying the roles played by experience and training in a system administrator’s ability to complete a set of tasks in a specified amount of time. In particular, she is interested in discriminating between administrators who are able to complete given tasks and those who are not. Data are collected on the performance of 75 randomly selected administrators. They are stored in the file SystemAdminstrations.csv. The variable Experience measures the months of full-time system administrator experience, while Training measures the number of relevant training credits. The dependent variable Completed is either Yes or No, according to whether or not the administrator completed the tasks. 1) Create a variable named isCompleted, which takes the value 1 if Completed == “Yes” and 0 otherwise. 2) Use the entire dataset as training data (i.e., no need to separate training data and validation data). Fit a logistic regression using isCompleted as the outcome variable, and Experience and Training as predictors. Report the following: a. The logistic regression model result (Put a screenshot of the regression summary and write out how the logistic regression model looks like). Logit(isCompleted) = [Intercept value] + [coefficient1]*Experience + [coefficient2]*Training b. Based on the result, does having more experience increase, decrease, or have no impact on completing the task? Why? c. Based on the result, does earning more training credits increase, decrease, or have no impact on completing the task? Why? 3) Provide the confusion matrix from the logistic regression using the cutoff value of 0.5 ( in other

words, an administrator is predicted to success if the predicted probability is higher than 0.5). You can put a screenshot of the results from R. 4) Based on the regression results, among those who successfully completed the task, what is the percentage of administrators who are incorrectly classified as failing to complete the task? Show the calculation steps that you use to answer this question.

5) If we want to decrease the percentage in question (4), should the cutoff level be increased or decreased? Why? Extra credit: Use the prediction model you obtain in question (2) and cutoff value of 0.5. At least how much experience (measured in month) should a programmer with 4 training credits have in order to be classified as being able to complete the task? Show the calculation steps that you use to answer this question. (Hint: Logit(0.5) = 0)

Submission checklist [ ] Answers to questions 2-5 and extra credit question; [ ] Screenshots of corresponding reports/charts for the above answers....