How to plot linear, cubic and quadratic fitting curve in a figure?

[Kristi Glover]

Hi R experts,
Could you please help me to fit a linear, cubic and quadratic curve in a figure? I was trying to show all these three fitting curves with different colour in one figure.
I spent substantial time to figure it out, but I could not.

I have given here a example and what I did for linear, but no idea for cubic and quadratic fitting curve


> dput(test)
structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame", row.names = c(NA,
-5L))
> plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20), ylim=c(0,14), ylab="S",xlab="env")
> linear<-lm(test$sp~test$env)
> quadratic<-lm(test$sp~test$env+I(test$env^2))
> #summary(quadratic)
> cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
> #summary(cubic)
> #fitting curve
> abline(linear)
>
Here I did for linear, but now I don't how I can plot quadratic and cubic line in the figure with different colour. I further wanted to put r2 value on the top of the fitting line.
ON linear fitting curve, I can see the line originated from left to right (cover whole x axis). I could not plot the line only within the data set.
Would any one help me to figure it? I think it is not difficult but for me- it is really taking time.

Thanks and waiting for your suggestions

sincerely,
Kristi Glover


 
     
        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
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.

[Kenneth Frost]

Hi, Kristi-
Here is an option.
?curve...to add the polynomials to your graph
?text....for adding the R2 to you plot
Ken


On 06/13/12, Kristi Glover  wrote:

>
> Hi R experts,
> Could you please help me to fit a linear, cubic and quadratic curve in a figure? I was trying to show all these three fitting curves with different colour in one figure.
> I spent substantial time to figure it out, but I could not.
>
> I have given here a example and what I did for linear, but no idea for cubic and quadratic fitting curve
>
>
> > dput(test)
> structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
> 20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame", row.names = c(NA,
> -5L))
> > plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20), ylim=c(0,14), ylab="S",xlab="env")
> > linear<-lm(test$sp~test$env)
> > quadratic<-lm(test$sp~test$env+I(test$env^2))
> > #summary(quadratic)
> > cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
> > #summary(cubic)
> > #fitting curve
> > abline(linear)
> >
> Here I did for linear, but now I don't how I can plot quadratic and cubic line in the figure with different colour. I further wanted to put r2 value on the top of the fitting line.
> ON linear fitting curve, I can see the line originated from left to right (cover whole x axis). I could not plot the line only within the data set.
> Would any one help me to figure it? I think it is not difficult but for me- it is really taking time.
>
> Thanks and waiting for your suggestions
>
> sincerely,
> Kristi Glover
>
>
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] mailing list
> 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.

______________________________________________
[hidden email] mailing list
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.

[Robert Baer]

> dput(test)
structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame", row.names =
c(NA,
-5L))

> plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20), ylim=c(0,14),
> ylab="S",xlab="env")
> linear<-lm(test$sp~test$env)
> quadratic<-lm(test$sp~test$env+I(test$env^2))
> #summary(quadratic)
> cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
> #summary(cubic)
> #fitting curve
> abline(linear)
>

Thanks and waiting for your suggestions

sincerely,
Kristi Glover

Try adding the following lines of code
cq = coef(quadratic)
cc = coef(cubic)
newenv = seq(min(test$env), max(test$env), by = (max(test$env) -
min(test$env))/500)
sp.quad = cq[1] + cq[2]*newenv +cq[3]*newenv^2
lines(newenv,sp.quad, col='red')

sp.cubic = cc[1] + cc[2]*newenv +cc[3]*newenv^2 +cc[4]*newenv^3
lines(newenv, sp.cubic, col='blue', lty=2)

------------------------------------------
Robert W. Baer, Ph.D.
Professor of Physiology
Kirksville College of Osteopathic Medicine
A. T. Still University of Health Sciences
800 W. Jefferson St.
Kirksville, MO 63501
660-626-2322
FAX [phone]

______________________________________________
[hidden email] mailing list
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.

[William Dunlap]

It is more direct to use predict() instead of reconstructing
by hand the prediction expression from the formula given
to lm().  E.g.,

  > x <- seq(1,6,by=1/4)
  > y <- sin(x) + rnorm(length(x), 0, 1/4)
  > plot(x, y)
  > fits <- lapply(1:3, function(degree)lm(y~poly(x, deg=degree)))
  > xpred <- pretty(x, n=50)
  > predictions <- lapply(fits, predict, newdata=list(x=xpred))
  > invisible(lapply(seq_along(fits), function(i)lines(xpred, predictions[[i]], col=i)))

If you change your fitting function, say to rq, or your formula, say to use a
categorical variable or interaction, you don't have to change anything else,
as the predict method for a model type takes care of the details.

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com

> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On Behalf
> Of Robert Baer
> Sent: Wednesday, June 13, 2012 8:40 AM
> To: Kristi Glover; R-help
> Subject: Re: [R] How to plot linear, cubic and quadratic fitting curve in a figure?
>
>
> > dput(test)
> structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
> 20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame", row.names =
> c(NA,
> -5L))
> > plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20), ylim=c(0,14),
> > ylab="S",xlab="env")
> > linear<-lm(test$sp~test$env)
> > quadratic<-lm(test$sp~test$env+I(test$env^2))
> > #summary(quadratic)
> > cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
> > #summary(cubic)
> > #fitting curve
> > abline(linear)
> >
> Thanks and waiting for your suggestions
>
> sincerely,
> Kristi Glover
>
> Try adding the following lines of code
> cq = coef(quadratic)
> cc = coef(cubic)
> newenv = seq(min(test$env), max(test$env), by = (max(test$env) -
> min(test$env))/500)
> sp.quad = cq[1] + cq[2]*newenv +cq[3]*newenv^2
> lines(newenv,sp.quad, col='red')
>
> sp.cubic = cc[1] + cc[2]*newenv +cc[3]*newenv^2 +cc[4]*newenv^3
> lines(newenv, sp.cubic, col='blue', lty=2)
>
> ------------------------------------------
> Robert W. Baer, Ph.D.
> Professor of Physiology
> Kirksville College of Osteopathic Medicine
> A. T. Still University of Health Sciences
> 800 W. Jefferson St.
> Kirksville, MO 63501
> 660-626-2322
> FAX [phone]
>
> ______________________________________________
> [hidden email] mailing list
> 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.

______________________________________________
[hidden email] mailing list
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.

[Frank Harrell]

See also the scripts in http://biostat.mc.vanderbilt.edu/BioMod

Frank
 

William Dunlap wrote

It is more direct to use predict() instead of reconstructing
by hand the prediction expression from the formula given
to lm().  E.g.,

  > x <- seq(1,6,by=1/4)
  > y <- sin(x) + rnorm(length(x), 0, 1/4)
  > plot(x, y)
  > fits <- lapply(1:3, function(degree)lm(y~poly(x, deg=degree)))
  > xpred <- pretty(x, n=50)
  > predictions <- lapply(fits, predict, newdata=list(x=xpred))
  > invisible(lapply(seq_along(fits), function(i)lines(xpred, predictions[[i]], col=i)))

If you change your fitting function, say to rq, or your formula, say to use a
categorical variable or interaction, you don't have to change anything else,
as the predict method for a model type takes care of the details.

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com


> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On Behalf
> Of Robert Baer
> Sent: Wednesday, June 13, 2012 8:40 AM
> To: Kristi Glover; R-help
> Subject: Re: [R] How to plot linear, cubic and quadratic fitting curve in a figure?
>
>
> > dput(test)
> structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
> 20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame", row.names =
> c(NA,
> -5L))
> > plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20), ylim=c(0,14),
> > ylab="S",xlab="env")
> > linear<-lm(test$sp~test$env)
> > quadratic<-lm(test$sp~test$env+I(test$env^2))
> > #summary(quadratic)
> > cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
> > #summary(cubic)
> > #fitting curve
> > abline(linear)
> >
> Thanks and waiting for your suggestions
>
> sincerely,
> Kristi Glover
>
> Try adding the following lines of code
> cq = coef(quadratic)
> cc = coef(cubic)
> newenv = seq(min(test$env), max(test$env), by = (max(test$env) -
> min(test$env))/500)
> sp.quad = cq[1] + cq[2]*newenv +cq[3]*newenv^2
> lines(newenv,sp.quad, col='red')
>
> sp.cubic = cc[1] + cc[2]*newenv +cc[3]*newenv^2 +cc[4]*newenv^3
> lines(newenv, sp.cubic, col='blue', lty=2)
>
> ------------------------------------------
> Robert W. Baer, Ph.D.
> Professor of Physiology
> Kirksville College of Osteopathic Medicine
> A. T. Still University of Health Sciences
> 800 W. Jefferson St.
> Kirksville, MO 63501
> 660-626-2322
> FAX [phone]
>
> ______________________________________________
> [hidden email] mailing list
> 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.

______________________________________________
[hidden email] mailing list
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.

[David Winsemius]

On Jun 13, 2012, at 10:29 AM, Kristi Glover wrote:

>
> Hi R experts,
> Could you please help me to fit a linear, cubic and quadratic curve  
> in a figure? I was trying to show all these three fitting curves  
> with different colour in one figure.
> I spent substantial time to figure it out, but I could not.
>
> I have given here a example and what I did for linear, but no idea  
> for cubic and quadratic fitting curve
>
>
>> dput(test)
> structure(list(sp = c(4L, 5L, 9L, 12L, 14L), env = c(12L, 18L,
> 20L, 17L, 15L)), .Names = c("sp", "env"), class = "data.frame",  
> row.names = c(NA,
> -5L))
>> plot(test$sp~test$env, main = "S vs. temp", xlim=c(0,20),  
>> ylim=c(0,14), ylab="S",xlab="env")
>> linear<-lm(test$sp~test$env)
>> quadratic<-lm(test$sp~test$env+I(test$env^2))
>> #summary(quadratic)
>> cubic<-lm(test$sp~test$env+I(test$env^2)+I(test$env^3))
>> #summary(cubic)
>> #fitting curve
>> abline(linear)
>>
> Here I did for linear, but now I don't how I can plot quadratic and  
> cubic line in the figure with different colour. I further wanted to  
> put r2 value on the top of the fitting line.
> ON linear fitting curve, I can see the line originated from left to  
> right (cover whole x axis). I could not plot the line only within  
> the data set.
> Would any one help me to figure it? I think it is not difficult but  
> for me- it is really taking time.
>

There is the poly  function that does the heavy lifting and also  
preserves the statistical propriety of polynomial fits.


 > quadratic<-lm( sp ~ poly(env, 2), data=test)
 > #summary(quadratic)
 > cubic<-lm( sp ~ poly(env, 3), data=test)
 > #summary(cubic)
 > #fitting curve

 > plot( sp~ env, data=test, main = "S vs. temp", xlim=c(0,20),  
ylim=c(0,14), ylab="S",xlab="env")

 > abline(linear)
 > lines( test$env[order(test$env)] , predict(quadratic)[order(test
$env)] )
 > lines( test$env[order(test$env)] , predict(cubic)[order(test
$env)] , col="red")

You can substitute:

lines(  x=seq( min(test$env) , max(test$env), len=100) ,
  y= predict(model, data.frame(env=seq( min(test$env) , max(test$env),  
len=100) ) ) )

... if you want smoother output.

--
David Winsemius, MD
West Hartford, CT

______________________________________________
[hidden email] mailing list
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.