[R] nls problem

[Rui Barradas]

Hello,

Sorry, disregard my previous e-mail.
Instead of your FPG function try


FPG <- function(pH, model) {
   coef(model)[1] + coef(model)[2]*pH + coef(model)[3]*pH^2
}

FPG(aedf$pH, m)
fitted(m)


Hope this helps,

Rui Barradas

Às 12:30 de 02/04/20, Rui Barradas escreveu:
> Hello,
> 
> Tip: in a formula, pH^2 = pH*pH is an interaction.
> 
> pH^2 = pH*pH = pH + pH + pH:pH = pH
> 
> Try I(pH^2)
> 
> Hope this helps,
> 
> Rui Barradas
> 
> Às 12:07 de 02/04/20, Troels Ring escreveu:
>> Dear friends - I'm on Win10 with R 6.3.1 and have a very simple 
>> problem with
>> nls which apparently gives a decent fit to the parable below, even 
>> without
>> starting values. But when I then think I know the meaning of the three
>> parameters a, b, and d it goes very wrong. I guess I am again overlooking
>> something easy but cannot spot it.
>>
>> BW
>> Troels Ring,
>>
>> Aalborg, Denmark
>>
>>
>>
>> aedf <- structure(list(Flux = c(-0.141256, -0.154709, -0.215247, 
>> -0.302691,
>>
>>              -0.32287, -0.511211, -0.605381, -0.813901, -1.11659, 
>> -1.76906
>>
>> ), pH = c(7.06273, 7.11182, 7.16182, 7.18818, 7.21455, 7.23818,
>>
>>            7.26273, 7.28455, 7.31182, 7.34364)), class = "data.frame",
>> row.names = c(NA,
>>
>> -10L))
>>
>> m <- with(aedf,nls(Flux~a + b*pH + d*pH^2))
>>
>>
>> with(aedf,plot(pH,Flux))
>>
>> with(aedf,lines(pH,fitted(m),lty=1,col="red",lwd=3))
>>
>>
>> m
>>
>> # a        b        d
>>
>> # -1630.70   457.67   -32.11
>>
>>
>> fitted(m)
>>
>> # 1] -0.22087053 -0.09989926 -0.13579690 -0.21936385 -0.34761742 
>> -0.50048799
>>
>> # [7] -0.69729602 -0.90471194 -1.20692552 -1.61994657
>>
>>
>> FPG <- function(pH) -1630.70 + 457.67*pH -32.11*pH^2
>>
>>
>> FPG(aedf$pH)
>>
>> # [1] -0.016359649  0.107602395  0.074773375 -0.007166685 -0.133786467
>>
>> # [6] -0.285187665 -0.480463769 -0.686513537 -0.987013685 -1.398026917
>>
>>
>> # So why aren't fitted(m) and FPG(aedf$pH) not closer ("equal")?
>>
>>
>>
>> This email has been scanned by BullGuard antivirus protection.
>> For more info visit www.bullguard.com
>> <http://www.bullguard.com/tracking.aspx?affiliate=bullguard&buyaffiliate=smt 
>>
>> p&url=/>
>>
>>     [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide 
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
> 
> ______________________________________________
> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.