In this example, the equation you get is y = a + bX = 13.932159 + 0.000698X.where the non negative constraint is UNchecked and you use GRG Nonlinear.Set Objective: you need to find the Minimum SSE figure … $G$7.For OLS, the R square value is found this way: =RSQ(B4:B13,A4:A13), in cell G8.You MUST find the SSE now, which is the sum of squared errors which means find the residuals for all rows and square each one then add all of them together. In cells G4 and G5 enter your initial guess … try to make an intelligent guess but don’t worry if you are wrong because SOLVER will find the right answer for you. This means, cell C4 contains the formula =G$4+G$5*A4 which you fill down to C13. Set up the worksheet as you see below and program the Predict Y column to feed off the range G4:G5. Linear Regression: Ordinary Least Squares … X v Y The best thing to do is to illustrate my ideas with examples. To use the SOLVER method you set up your model and enter estimates for the answers you need. Jut when you think it’s a waste of time to learn yet another regression technique, SOLVER will solve your simple regression problems, your logarithmic, power, exponential and polynomial problems. To prove that, I build a series of models using SOLVER and found that it is true. I was just reading about things in general when something just dawned on me: SOLVER will solve regression problems.
0 kommentar(er)
