Linear Algebra is Your Friend: Least Squares Solutions to Overdetermined Linear Systems

Most engineering students are quite familiar with least squares regression analysis of experimental data on Y versus X plots using EXCEL or other similar spreadsheet software programs.They use the reported coefficient of determination, R 2 , as a measure of the "closeness of fit".Most chemical engineering students, however, are not well versed enough with the concepts of modern day computational techniques for finding the least squares solution or with the application of linear algebra for determining the best solution to overdetermined linear systems of equations.