Matrix dimnames different from sequence of values

Question

I have this code to create a matrix:

TTCm <- seq(from=-4, to=0, by=0.1)
Speedm <- seq(from=15, to=25, by=0.5)
matrix2 <- matrix(nrow = 41, ncol = 21)
dimnames(matrix2) = list(TTCm, Speedm)

However (depending on the sequence) dimnames might give weird values. In the above particular case TTCm -0.1 is -0.0999999999999996. Does anyone know what the problem here is? Thanks

https://stackoverflow.com/questions/588004/is-floating-point-math-broken — Anger Density, Aug 22 '19 at 12:23
Also read: https://stackoverflow.com/questions/9508518/why-are-these-numbers-not-equal — Ronak Shah, Aug 22 '19 at 13:24

BBang · Accepted Answer · 2019-08-22T13:33:48.023

0

That's because computer uses finite precision arithmetic. it's well known as "floating point problem". so the 40th value of TTcm you can see on R console is a approximate value. you can check it by using dplyr::near() on R console. Try the code below.

TTCm <- seq(from=-4, to=0, by=0.1) TTCM[40] == -0.1 dplyr::near(TTCm[40], -0.1)

if you don't understand completely, Try the code below. Maybe you can totally understand.

sqrt(2) ^ 2 == 2 1/10 * 10 == 1 near(sqrt(2)^2, 2) near(1/10*10, 1)

you need to be careful when handling floating point.

edited Aug 22 '19 at 13:33

answered Aug 22 '19 at 12:47

BBang

16
2

And what would be the easiest way to get around this? thanks – Attila Borsos Aug 22 '19 at 12:59
in the case, the easiest way is.. Try it like this. `dimnames(matrix2) =list(round(TTCm,10), Speedm)` I don't know this is the answer you wanted.. – BBang Aug 22 '19 at 13:08

Matrix dimnames different from sequence of values

1 Answers1