Skip to content

how to plot a logarithmic regression line

2 messages · arun, David Winsemius

arun
# 
HI,
Try ?curve

fit <- lm(Mean_Percent_of_Range~log(No.ofPoints))
?coef(fit)
?# ?? (Intercept) log(No.ofPoints) 
? # ??? -74.52645???????? 46.14392 



?plot(Mean_Percent_of_Range ~ No.ofPoints) 
curve(coef(fit)[[1]]+coef(fit)[[2]]*log(x),add=TRUE,col=2)


A.K.



I realize this is a stupid question, and I have honestly tried to find 
the answer online, but nothing I have tried has worked. I have two 
vectors of data: 

"Mean_percent_of_range" 
10.90000 ?17.50000 ?21.86667 ?25.00000 ?25.40000 ?26.76667 ?29.53333
 ?32.36667 ?43.13333 ?41.80000 50.56667 ?49.26667 ?50.36667 ?51.93333 
?59.70000 ?63.96667 ?62.53333 ?60.80000 ?64.23333 ?66.00000 74.03333 
?70.40000 ?77.06667 ?76.46667 ?78.13333 ?89.46667 ?88.90000 ?90.03333 
?91.60000 ?94.30000 95.50000 ?96.20000 ?96.50000 ?91.40000 ?98.20000 
?96.60000 ?97.40000 ?99.00000 100.00000 

and 
"No.ofPoints" 
5 ?6 ?7 ?8 ?9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 
39 40 41 42 43 

When I plot these, I get a logarithmic curve (as I should for this type of data)
All that I want to do is plot best fit regression line for that 
curve. From what I have read online, it seems like the code to do that 
should be
but that gives me a straight line that isn't even close to fitting the data 

How do I plot the line and get the equation of that line and a correlation coefficient? 
Thanks
David Winsemius
#
On Feb 22, 2014, at 1:06 PM, arun wrote:

            

      
      
      
    

        
The 'abline' function is not what you want. Use 'lines' to plot multiple points. 

Perhaps:

mod <- lm(log(Mean_percent_of_range) ~ log(No.ofPoints))
 plot(log(Mean_percent_of_range), log(No.ofPoints))
lines( log(No.ofPoints), predict(mod))
#------------

            

      
      
      
    

        
Call:
lm(formula = log(Mean_percent_of_range) ~ log(No.ofPoints))

Residuals:
     Min       1Q   Median       3Q      Max 
-0.32617 -0.04839  0.00962  0.05316  0.17316 

Coefficients:
                 Estimate Std. Error t value Pr(>|t|)    
(Intercept)       1.19840    0.08060   14.87   <2e-16 ***
log(No.ofPoints)  0.94228    0.02609   36.12   <2e-16 ***
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1

Residual standard error: 0.09455 on 37 degrees of freedom
Multiple R-squared:  0.9724,	Adjusted R-squared:  0.9717 
F-statistic:  1305 on 1 and 37 DF,  p-value: < 2.2e-16



David Winsemius
Alameda, CA, USA