significance of non-linear regression parameters
Message posted by Mike Yong (via 161.252.32.91) on May 14, 2001 at 6:15 AM (ET)

Hello listers,

Suppose I have a non-linear equation, y = a (1-exp(b+cx-d sin((x-e)), I can use least sqaures method to fit the equation and derive estimates for a, b, c,d, and e. However, if the t value for some paprameter is very low, say 0.2, how should explain the result? Should I still keep that parameter? If the significance of non-linear regression parameters can be explained as linear equation parameters, the parameter having a low t vlaue should be droped and the equation should be revised. Am I right?
I thank you so much in advance for any potential help.