Verified ISYE 6501 Final Quiz - Summer 2021

Document Content and Description Below

Drag each of the solution methods to the type of problem it's designed for. Each model or method should be placed somewhere, so you may need to put more than one in the same answer box. There may be m... ore than one correct answer for some models/methods. Drag each of the models to the type of data and type of question it's best suited for. Each model or method should be placed somewhere, so you may need to put more than one in the same answer box. There may be more than one correct answer for some models/methods. Select all of the following that are examples of time-series data Select all of the following reasons that data should not be scaled until point outliers are removed. Select all of the following situations in which using a variable selection approach like lasso or stepwise regression would be important. Drag each of these software packages to the model it's best suited for analyzing. Each model or method should be placed somewhere, so you may need to put more than one in the same answer box. For each of the analytics tasks listed below, drag to it the R function(s) that do it. Not all functions might be used. The following process was followed to predict sales of a product each month for the next three years: A positive correlation has been observed between hours of sleep and self-reported happiness. Select all of the following situations where imputing missing data for a variable is probably better than than including a "data missing" binary variable? Select all of the following situations where a linear regression model is more directly appropriate than a logistic regression model when analyzing a specific World Cup soccer match. Select all of the following situations where a supervised learning model (like classification) is more directly appropriate than an unsupervised learning model (like clustering). A hospital has collected data on how long hip replacement surgery patients have required before regaining nearly-full motion without pain, as well as attributes of each patient (age, height, weight, pre-surgery range of motion, other medical conditions, etc.). Now, the hospital wants to use that data to predict recovery time for a new patient. Select all of the following situations where a simulation model is more directly appropriate than an optimization model. A large trial law firm would like to increase the fraction of cases it wins by doing a better job of assigning cases to its lawyers. If you're an expert in the legal industry, please do not rely on your expertise to fill in all that extra complexity (you'll end up making the questions more complex than I intended). A support vector machine model has been created to predict whether a person is right-handed or left-handed, based on the person's genetic profile. The figure above shows a confusion matrix of the model's performance on a test data set that it was not trained on. very large (thousands of rooms) hotel in Las Vegas is planning to remodel its parking garage. There are lots of considerations and complexities that go into doing that, so this question will look at just a small part of it, with several simplifications. If you're an expert in the construction or hotel industries, please do not rely on your expertise to fill in all that extra complexity (you'll end up making the questions more complicated than I intended). When remodeling the parking garage, the hotel wants to make sure that it is unlikely to run out of parking spaces even when all rooms are occupied, but given that restriction it also wants to make the parking garage as small as possible to save costs and space. The hotel would like to use analytics (analyzing its past ten years of data) to help determine the right number of parking spaces to have. A complicating factor is that the hotel doesn't have complete data; for about 2% of the hotel guests, the person at the front desk did not record whether the guest had a car or not. a. The hotel's director of facilities has come up with the following incorrect idea: GIVEN past attribute data of guests and whether they had a car, USE linear regression TO impute the missing data (whether guests had a car or not). Then, GIVEN the number of cars each day for the past ten years, USE exponential smoothing TO predict how many parking spaces will be needed each day for the next ten years. Finally, GIVEN the daily predictions of the number of parking spaces required, USE optimization TO determine the minimum number of parking spaces required so that 90% of the highest 10% of daily predictions will be less than or equal to the number of parking spaces. Select all of the statements below that show a reason why the director's idea is wrong. b. Select a set of models from the list below, that the director can put together to determine how many parking spaces there should be. c. Select all of the following complexities that are not accounted for in any of the models in part d. n the United States in 2016, the median annual earnings among working women were about 20% lower than the median annual earnings among working men. a. What nonparametric test could most-directly be used to show that the difference between the median incomes of working men and working women is statistically significant? In the United States in 2016 the fraction of working men who died on the job was about 11 times higher than the fraction of working women who died on the job. b. Select all of the appropriate uses of the binomial distribution to test whether the difference in death rates between working men and working women is statistically significant. Let Nm be the number of working men, Nw be the number of working women, and Km and Kw be the number of men and women killed on the job. NOTE: In the answer choices below, a "yes" answer refers to what is usually called a "success" in the binomial distribution, but obviously a person being killed is not a "success" so I've used a different term. c. Which of the nonparametric tests that would be valid to use in such a study. [Do not spend time worrying about whether the study's setup provides a valid comparison For the sake of this question assume it does and think about what nonparametric test could be used in that case.] Finding pairs of people like the ones suggested in part c. could be very difficult. Instead, the researcher might want to use a different type of model. d. Select all of the approaches below that might help determine how much of the differences remain after accounting for all of the factors listed above in part c. Do you think that you or any of your fellow students in this course would be good TAs for the course in the future? If so, please enter name(s) or username(s) below. [Show More]

Last updated: 1 year ago

Preview 1 out of 20 pages