Materials for Seminar

Gion Matsuri

Materials for Seminar

July 19th
Title: Regression as a Special Case of Quadratic Programming

Students in operations research and industrial engineering typically study linear and non-linear programming. Whereas regression is more commonly used in the fields of statistics and econometrics. This lecture will describe the relationship between the two methodologies. Specifically, standard ordinary least squares regression can be formulated as a quadratic programming problem and solved by finding a solution to a linear set of equations. I will demonstrate how to formulate and solve regression problems in a number of ways using the software GAMS. A temporary license for GAMS will be provided and students should come to class with GAMS already installed on their computers.


Directions for Installing GAMS

If you are having trouble installing GAMS, please uninstall GAMS completely. Then use the directions I have provided below. I used these directions on a computer with Japanese Windows and had no problems installing GAMS.

  • Download license file from email
  • Download software from gams website
  • Double click on downloaded file to begin installation
  • Click next twice
  • Make sure “Use GAMS IDE” is selected and click next
  • Click install
  • After the files finish installing, click the radio button next to “Copy license file” and click finish
  • Browse to the location of your license file, highlight it, then click open at the bottom

Least Squares

OLS in GAMS (Primary): We will cover pages 1-13

Regression in GAMS:
Klein Example pages 44-52

GAMS code for Regression: Example
Example Confidence Intervals

Assignment for lecture

Data generation file (New!)
Example updated ghj.gms file (New!)

July 19th Lecture
July 26th Lecture