Section-3

.docx

School

University of Texas, San Antonio *

*We aren’t endorsed by this school

Course

ECONOMETRI

Subject

Economics

Date

Jan 9, 2024

Type

docx

Pages

4

Uploaded by Jholmesacts

Exam PA – Section 3.4 Question 3 Business Problem You are a consultant for a regional hospital outside of a major metropolitan area. Due to growing concerns of the hospital nearing full capacity, the hospital has come to your consulting firm to determine the number of visits a patient is likely to make in the upcoming 2 weeks given the medical information they are allowed to provide, which is heavily restricted due to the Health Insurance Portability and Accountability Act. Note that the hospital is strongly concerned with interpretation. Doctors feel that they should be able to communicate the results to hospital administrators. Specific Task You decide that the best course of action is to model the data using GLM. To kick things off, you investigate the quality of the provided dataset. However, the time sensitive nature of this project means you are not able to spend time fixing or cleaning any errors or issues that might be present in the data. In carrying out due diligence, you choose to at least note any concerns you find. (a) Comment on your findings. ANSWER: Your supervisor recommends the gaussian family with identity link to model the number of hospital visits. However, you decide to proceed with the Poisson family and log link. (b) Critique your supervisor’s recommendation in light of your modeling choices. ANSWER:
You decide to run a stepwise selection procedure to determine what features should belong in your GLM. (c) (i) Recommend a specific stepwise procedure. Justify your recommendation. (ii) Provide details on the coefficients of both the full GLM and the resulting stepwise model. ANSWER: Code is provided to obtain log-likelihoods in order to assess the two models that have been created. You consider it as part of your analysis in choosing a final model. (d) Recommend your model of choice. Justify your recommendation. ANSWER: Code is provided to aid with interpreting your chosen model from (d) by exponentiating its coefficient estimates. (e) (i) Explain the reason for exponentiating the estimates. (ii) List several of the most impactful predictors. (iii) Describe the impact of each predictor mentioned in (ii) in terms of model predictors. ANSWER:
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help