how to plot a logarithmic regression line
On Feb 22, 2014, at 1:06 PM, arun wrote:
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)
plot(Mean_Percent_of_Range ~ No.ofPoints)
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
abline(lm(log(Mean_Percent_of_Range) ~ log(No.ofPoints)))
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?
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)) #------------
summary(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