Linear Programming (R, Python) Assignment Example, Singapore

Programming Assignment

Instructions

• You are allowed to use the following software to answer the questions:

– R and Rstudio

– Python

• Some questions are designed to be open-ended and have several correct answers or multiple ways to express your answer. Strive to be concise and precise in your answers. Excessive words or quotes will be penalized. You are notexpected to write essay, and excessive words signal that you do not know what is important.
• You may use more than one page per question, if necessary, but start each new question number on a new page.

Hospital Length of Stay Estimation at Admission

Inpatient Length of Stay (LOS) in hospitals are important estimates done upon admission (after a tentative diagnosis) to plan for hospital resources (beds, specialized equipment, doctors/nurses with specific training, etc…) and to inform the patient on the estimated cost of stay so that they can get the required approval from insurance. Currently, the estimation is done using the mean LOS in each CCSR diagnosis.

A data sample (INF002v4.csv) and data dictionary from the New York State Department of Health are provided. The length of stay value of 120+ had been recoded to 120 so that you can use LOS as a continuous integer variable.
The business objective is to estimate the LOS more accurately upon admission than using the mean LOS.

1. Explore the data and report 3 notable findings.

2. State the list of potential predictor X variables that will be inputs to models and the final dimensions of the dataset (i.e. number of observations and variables).

3. Using 70-30 train-test, conduct (a) Linear Regression, and (b) CART to compare testset errors. Explain your specification of the model1 and display the results in a table.

4. Comment on your models’ results and provide insights for the business application.

5. Suggest improvements.