Origin Version (7.0 or 7.5): Operating System:XP someone does multipke regression with data showed below; A[Y] B[X1] C[X2] 17 1 10 24 2 20 31 3 30 38 4 40 45 5 50 the result should be:Y=10+2*X1+0.5*X2,but the origin gives nothing, all parameters are 0, is it a bug of origin on multiple regression?
I'll have to leave it to someone with a better grasp of multiple regression theory to explain just why this fails, but I would suggest that your data aren't appropriate for this type of test. This, I believe, is related to the fact that the relationship between your dependent and each of the two independent variables is perfectly linear -- hence there is no deviation and no error to be calculated.
I tried your dataset with Excel's regression tool and it failed, too. Fudging your data ever so slightly so that the relationships were no longer perfectly linear produced similar results in both Origin and Excel.
So, I think that if there is a problem in Origin, it is that we don't report that the data aren't appropriate to selected analysis. We should probably fix that.
Is it a bug about multipke regression?
Posted - 06/12/2005 : 11:54:13 PM
