In the R expression
x[1] + x[2]
the result must be stored as a double precision number,
because that is what R "numerics" are. sum() does not
have to keep its intermediate results as doubles, but
can use quad precision or Kahan's summation algorithm
(both methods involve more than a simple doub
This is described in R FAQ 7.31
ir. Thierry Onkelinx
Statisticus / Statistician
Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkel...@inbo.
On Sun, 25 Feb 2018, Iuri Gavronski wrote:
Hi,
Why sum() on a 10-item vector produces a different value than its
counterpart on a 2-item vector? I understand the problems related to
the arithmetic precision in storing decimal numbers in binary format,
but shouldn't the errors be equal regardles
Hi,
Why sum() on a 10-item vector produces a different value than its
counterpart on a 2-item vector? I understand the problems related to
the arithmetic precision in storing decimal numbers in binary format,
but shouldn't the errors be equal regardless of the method used?
See my example:
> opti
On Jan 15, 2009, at 12:25 PM, Charles C. Berry wrote:
This is what I get on windows XP:
tcp1-tcp2
[,1] [,2] [,3]
[1,] -2.861023e-06 -4.768372e-07 -4.768372e-07
[2,] -4.768372e-07 -3.814697e-06 2.622604e-06
[3,] -4.768372e-07 2.622604e-06 -5.960464e-08
bu
On 1/15/2009 1:32 PM, David Winsemius wrote:
On Jan 15, 2009, at 12:25 PM, Charles C. Berry wrote:
This is what I get on windows XP:
tcp1-tcp2
[,1] [,2] [,3]
[1,] -2.861023e-06 -4.768372e-07 -4.768372e-07
[2,] -4.768372e-07 -3.814697e-06 2.622604e-06
[3,] -4.
Hi there,
Thanks for the help.
I see now where my results are coming from.
Marlon.
-Original Message-
From: Charles C. Berry [mailto:cbe...@tajo.ucsd.edu]
Sent: Friday, 16 January 2009 6:26 a.m.
To: dos Reis, Marlon
Cc: David Winsemius; r-help@r-project.org
Subject: RE: [R] Precision in
ited accuracy of floating point calcs on the system used.
HTH,
Chuck
Marlon.
-Original Message-
From: David Winsemius [mailto:dwinsem...@comcast.net]
Sent: Thursday, 15 January 2009 6:04 p.m.
To: Charles C. Berry
Cc: dos Reis, Marlon; r-help@r-project.org
Subject: Re: [R] Precision in R
I
81608.66 109642.09
[3,] -78368.17 109642.09 -147305.32
solve(tcp2)
Error in solve.default(tcp2) :
system is computationally singular: reciprocal condition number =
2.17737e-17
Marlon.
-Original Message-
From: Charles C. Berry [mailto:cbe...@tajo.ucsd.edu]
Sent: Thursday, 15 January 2009 5:16
ent: Thursday, 15 January 2009 5:16 p.m.
To: dos Reis, Marlon
Cc: r-help@r-project.org
Subject: Re: [R] Precision in R
Marlon,
Are you using a current version of R? sessionInfo()?
It would help if you had something we could _fully_ reproduce.
Taking the _printed_ values you have below (WBtWB)
age-
From: David Winsemius [mailto:dwinsem...@comcast.net]
Sent: Thursday, 15 January 2009 6:04 p.m.
To: Charles C. Berry
Cc: dos Reis, Marlon; r-help@r-project.org
Subject: Re: [R] Precision in R
I am seeing different behavior than don Reis on my installation as well:
mtx is the same as his WBtW
I am seeing different behavior than don Reis on my installation as well:
mtx is the same as his WBtWB
> mtx <- matrix(c(1916061939, 2281366606, 678696067, 2281366606,
3098975504, 1092911209, 678696067, 1092911209, 452399849), ncol=3)
>
> mtx
[,1] [,2] [,3]
[1,] 1916061
Marlon,
Are you using a current version of R? sessionInfo()?
It would help if you had something we could _fully_ reproduce.
Taking the _printed_ values you have below (WBtWB) and adding or
subtracting what you have printed as the difference of the two visually
equal matrices ( say Delta ) ,
rom: Nathan S. Watson-Haigh [mailto:nathan.watson-ha...@csiro.au]
Sent: Thursday, 15 January 2009 3:25 p.m.
To: dos Reis, Marlon
Cc: r-help@r-project.org
Subject: Re: [R] Precision in R
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
I've recently been working with cross products etc.
You should try the follow
Yes, computing WB.%*%t(WB) may be the problem, by either method.
if the goal is to compute the inverse of WB%*%t(WB), you should
consider computing the singular value or QR decomposition for the
matrix WB.
If WB = Q%*%R, where Q is orthogonal, then WB %*% t(WB) =R %*%t(R),
and the inverse of
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
I've recently been working with cross products etc.
You should try the following function:
tcp1 <- tcrossprod(WB)
or
tcp2 <- crossprod(t(WB))
Both should be the same (check for equality accounting for some floating point
imprecision):
all.equal(tcp1
Dear All,
I'm preparing a simple algorithm for matrix multiplication for a
specific purpose, but I'm getting some unexpected results.
If anyone could give a clue, I would really appreciate.
Basically what I want to do is a simple matrix multiplication:
(WB) %*% t(WB).
The WB is in the disk so I com
17 matches
Mail list logo
