0

I made a non linear regression in R (package quantreg) and obtained the following result:

Coefficients:
     Value    Std. Error t value  Pr(>|t|)
Asym 30.98339  1.06558   29.07656  0.00000
mid  36.18334  1.49832   24.14935  0.00000
scal  1.75042  1.07709    1.62514  0.10573

Now, I want to understand what it means exactly. So, the first parameter is the asymptote, which I expect to be the value of y at which further increases of the predictive factor no longer leads to an increase in y values.

Then it comes the Mid, I asume it is some sort of inflexion point, or a value of X at which y starts to rapidly lose its increase.

Finally, the scaling factor scal, would show how much y increases when x increase before the inflexion point.

The standard error is easy to understand for all coefficients´ values, but the T is not so. I kinf of imagine that, for a scaling factor, the T means the increase in y divided by the error, but what would be the interpretation of a high T for an asymptote or an inflexion point?

Likewise, the probability of t being different from zero, seems straightforward to interpret, but what does it mean to have significant values of t for the asymptote and the inflexion?

I could not find answers to these questions elsewhere. Thanks in advance for any clarification

Agus camacho
  • 868
  • 2
  • 9
  • 24
  • The t-value is the T statistic for the individual coefficient significance tests. The probability (column to the right) is the p-value for that t-student distribution value. – Celius Stingher Oct 08 '19 at 21:20
  • 1
    Thank you, I know how they are calculated. My questions try to clarify the interpretation of these values. – Agus camacho Oct 10 '19 at 19:02

0 Answers0