Page 1 of 1
Regression small numbers
#1
Posted 17 August 2007 - 05:20 AM
I have a set of numbers, one side of the set is a larger number relative to the other one, but the changes to the one side are small relative to the other. Is it permissible to work with the residual to establish a trend line for the small changes?
21.720840 1.000058954
25.321081 1.000080116
29.518064 1.000108876
34.410700 1.000147960
40.114292 1.000201074
46.763259 1.000273254
54.514295 1.000371346
63.550070 1.000504649
74.083529 1.000685805
Can I subtract 1 and then strip off 3 decimals in the right column and work with that and the left column to establish a curve fit for the residual changes in the right column? Does this in anyway invalidate the curve fit?
21.720840 1.000058954
25.321081 1.000080116
29.518064 1.000108876
34.410700 1.000147960
40.114292 1.000201074
46.763259 1.000273254
54.514295 1.000371346
63.550070 1.000504649
74.083529 1.000685805
Can I subtract 1 and then strip off 3 decimals in the right column and work with that and the left column to establish a curve fit for the residual changes in the right column? Does this in anyway invalidate the curve fit?
Share this topic:
Page 1 of 1

Help
Join now

Promote to Article








