CODEWITHSUNDEEP
rnonlinear-functionsnatural-logarithm
wernerfeuer
wernerfeuer

Reputation: 575

R: Using equation with natural logarithm in nls

Good day,

I am struggling with R and natural logarithm (ln). Firstly, I cannot find a ln(x) function in R. I have noticed that log(x) is the same as ln(x) (when using ln(x) with a calculator).

In R:

log(5) = 1.609438

And with a calculator:

ln(5) = 1.609438
log(5) = 0.69897

I'm trying to fit an equation in R (this is exactly how I found in the literature of 3 references):

y = a + b(x/305) + c(x/305)2 + d ln(305/x) + f ln2(305/x)

Is it correct to use the following syntax in R to use the equation?

y ~ a + b*(x/305) + c*((x/305)^2) + d*log(305/x) + f*(log(305/x))^2

The idea is to use this function with nls() in R. Thanks in advance!

Upvotes: 22

Views: 131606

Answers (3)

alko989
alko989

Reputation: 7908

In R, log is the natural logarithm. In calculators, log usually means base 10 logarithm. To achieve that in R you can use the log10 function.

log(5)
## [1] 1.609438
log10(5)
## [1] 0.69897

As for your formula, it seems correct, since log is the natural logarithm.

Upvotes: 43

Jackson Thounaojam
Jackson Thounaojam

Reputation: 11

In R, log computes logarithms, by default natural logarithms, log10 computes common (i.e., base 10) logarithms, and log2 computes binary (i.e., base 2) logarithms. The general form log(x, base) computes logarithms with base.("R Documentation")

Upvotes: 1

user20637
user20637

Reputation: 674

In addition I will point out that your model

y ~ a + b*(x/305) + c*((x/305)^2) + d*log(305/x) + f*(log(305/x))^2

is linear in the statistical sense of being linear in the coefficients; it doesn't need to be linear in x.

You don't need nls to fit this model, you could use lm().

But remember to look at the I() function to express terms like (x/305)^2.

ETA example:

aDF <- data.frame(x=abs(rnorm(100)), y=rnorm(100))
lm(y ~ 1 + I(x/305) + I((x/305)^2) + log(305/x) + I(log(305/x)^2), data=aDF)

Upvotes: 4

Related Questions