Use `nlsfit` within geom_smooth to add exponential line to plot

问题

I would like to use the nlsfit from the easynls package with ggplot2 if at all possible.

This is what I have done so far:

Set up subset data:

library('ggplot2')
library('easynls')

x <- seq(25,97)
y <- c(0.014, 0.016, 0.015, 0.016, 0.018, 0.019, 0.023, 0.019, 0.021, 0.017, 0.018, 0.016, 0.016, 0.020, 0.018, 0.019, 0.022, 0.023, 0.027, 0.027, 0.028, 0.031, 0.029, 0.032, 0.030, 0.030, 0.030, 0.033, 0.039, 0.038, 0.039, 0.046, 0.042, 0.043, 0.050, 0.054, 0.059, 0.064, 0.062, 0.058, 0.063, 0.069, 0.071, 0.069, 0.073, 0.071, 0.070, 0.077, 0.086, 0.077, 0.090, 0.086, 0.098, 0.108, 0.112, 0.116, 0.129, 0.120, 0.128, 0.141, 0.150, 0.143, 0.148, 0.150, 0.162, 0.162, 0.168, 0.152, 0.151, 0.161, 0.169, 0.189, 0.184)
data <- data.frame(x,y)

Run NLSfit on sample data

nlsfit = nlsfit(data.frame(x,y), model=6, start=c(250,0.05))
nlsfit
# $Model
# [1] "y~a*exp(b*x)"

# $Parameters
#                              y
# coefficient a           0.0061
# coefficient b           0.0358
# p-value t.test for a    0.0000
# p-value t.test for b    0.0000
# r-squared               0.9793
# adjusted r-squared      0.9790
# AIC                  -500.0812
# BIC                  -493.2098

Plot using plot() with a line

plot(x, y)
a <- nlsfit$Parameters[1,]
b <- nlsfit$Parameters[2,]
lines(x, a*exp(x*b), col="steelblue")

Attempt to use nls with ggplot2 (this works - but the fit isn't as good on the full dataset)...

ggplot(data, aes(x=x, y=y)) + geom_point(
       ) + geom_smooth(method="nls", formula=y~a*exp(x*b),
       method.args=list(start=c(a=250,b=0.05)), se=FALSE)

Attempt to nlsfit with ggplot2 -- doesn't work

# Below doesn't work
ggplot(data, aes(x=x, y=y)) + geom_point(
       ) + geom_smooth(method="nlsfit", formula=y~a*exp(x*b),
       method.args=list(data.frame(x, y),
                        model=6, start=c(250,0.05)), se=FALSE)

# Warning message:
# Computation failed in `stat_smooth()`:
# unused arguments (formula, weights = weight, list(x = 25:97, y = c(0.014, 0.016, 0.015, 0.016, 0.018, 0.019, 0.023, 0.019, 0.021, 0.017, 0.018, 0.016, 0.016, 0.02, 0.018, 0.019, 0.022, 0.023, 0.027, 0.027, 0.028, 0.031, 0.029, 0.032, 0.03, 0.03, 0.03, 0.033, 0.039, 0.038, 0.039, 0.046, 0.042, 0.043, 0.05, 0.054, 0.059, 0.064, 0.062, 0.058, 0.063, 0.069, 0.071, 0.069, 0.073, 0.071, 0.07, 0.077, 0.086, 0.077, 0.09, 0.086, 0.098, 0.108, 0.112, 0.116, 0.129, 0.12, 0.128, 0.141, 0.15, 0.143, 0.148, 0.15, 0.162,
# 0.162, 0.168, 0.152, 0.151, 0.161, 0.169, 0.189, 0.184)))

Is this possible - would appreciate any help. Thanks.

回答1:

You can try stat_function to make the last part work:

a <- nlsfit$Parameters[row.names(nlsfit$Parameters) == 'coefficient a',]
b <- nlsfit$Parameters[row.names(nlsfit$Parameters) == 'coefficient b',]
ggplot(data, aes(x=x, y=y)) + geom_point() + 
  stat_function(fun=function(x) a*exp(b*x), colour = "blue")

来源：https://stackoverflow.com/questions/41881329/use-nlsfit-within-geom-smooth-to-add-exponential-line-to-plot

标签

ggplot2

machine-learning

nls