CODEWITHSUNDEEP
rggplot2curve-fittingcurveexponential
Leandro Romero
Leandro Romero

Reputation: 11

Exponential Fitting in R

I have a data set with two columns and I want to fit to an exponential curve and show the equation on the chart. Could you please help me with the code?

x<-1:6
y<-c(86000,114597,165576,207467,328745,531531)
df<-as.data.frame(cbind(x,y))

ggplot(df,aes(x,y))+
  geom_point()+
  geom_line(colour=" dark blue")+
  geom_smooth(method = "nls",se = F)+
  scale_y_continuous(labels = scales::comma)+
  theme_economist()+
  xlab("Grade")+
  ylab("Salary avg")

when I try to use nls I can not see the curve on ggplot.

Upvotes: 1

Views: 541

Answers (1)

Ben Bolker
Ben Bolker

Reputation: 226182

Setup:

library(ggplot2)
df <- data.frame(x=1:6,
      y=c(86000,114597,165576,207467,328745,531531))
gg0 <- ggplot(df,aes(x,y))+
  geom_point()+
  geom_line(colour=" dark blue")+
  scale_y_continuous(labels = scales::comma)+
  ggthemes::theme_economist()+
  xlab("Grade")+
  ylab("Salary avg")

One way to do it is to fit a GLM with Gaussian family and log link:

gg0 + geom_smooth(method="glm",
    formula=y~x,
    method.args=list(family=gaussian(link="log")))

You can use nls, but it's a little tricky. I used a self-starting method here, but (1) had to install an archived package, (2) had to specify start anyway (I think it's ignored but needs to be there)

## remotes::install_version("drLumi",version="0.1.2")
gg0 + geom_smooth(method="nls",
    formula=y~drLumi::SSexp(x,b,y0), 
     se=FALSE,
     method.args=list(start=list(b=1,y0=1)))

Incomplete attempt to extract regression equation:

g1 <- glm(y~x, data=df, family=gaussian(link="log"))
library(latex2exp)
library(equatiomatic)
s <- TeX(extract_eq(g1,ital_vars=TRUE))

Upvotes: 1

Related Questions