What do the zero values indicate in the regression coeffient output?

After using the following codes:

X = [ones(length(x1),1) x1' x2' x3'];

B = X\Y';

I obtain the following coefficients in a [2 x 4 matrix]:

0 0.006 0.0021 0.001

0 0.0046 0.0009 -0.0012

What do the zero values in these coefficent estimates indicate?

Answer 1

[Star Strider]

In the situation you describe, with a vector of ones in the first column of your X matrix corresponding to the parameter that equals zero, the zero value means that the offset (analogous to the y-intercept b in the linear equation y=m*x+b) is zero. The equation is essentially y=m*x.

Stated more simply, it means that the hyperplane that fits your data passes through the origin. Your Y data are likely described by x1, x2, and x3 only. However I suggest you use regress (Statistics Toolbox) to get the confidence intervals on the parameters (and optionally other statistics on the fit) before you draw any firm conclusions.

[Tim Bennett]

Thats great. Thanks for your help.

[Star Strider]

As always, it is my pleasure to help.