I was wondering if there is a counterpart to scipy.linalg.cho_solve in R. What the function does is given the cholesky factor L of A (A = LL') and b, it solves the original problem, Ax = b. (not Lx = b)
(So it is different from backsolve/forwardsolve)
Thank you,
Joon
I can't think of a function doing that for you automatically, but given you have the cholesky factor L, it's easily done in one line by reconstructing the A matrix as defined by the decomposition A=LL' :
A=matrix(c(1,1,1,1,5,5,1,5,14),nrow=3)
# Cholesky decomposition A = LL'
L <- chol(A)
# Make some b with known x
x <- c(1,2,3)
b <- A%*%x
# Solve
solve( t(L) %*% L, b)
edit: be aware that in R, the definition of the Cholesky factor is related to A = L'L, which is why you have to put the transposed first in the solve.
edit2 : After reading Bates article, I realized it should be:
> solve(crossprod(L),b)
[,1]
[1,] 1
[2,] 2
[3,] 3
If I understand you correctly, then Doug Bates covered some of this in an article he wrote for R News in 2004 (see page 18 of the linK).
The relevant bit is:
ch <- chol(crossprod(X))
chol.sol <- backsolve(ch, forwardsolve(ch, crossprod(X, y),
upper = TRUE, trans = TRUE))
where X is the matrix of predictor variables.
Doug's article goes on to show how functionality in the Matrix package (which comes with R) can be used solve the same system very quickly indeed.
I realise this question is a little old, but I see that the answer
forwardsolve(L, forwardsolve(L, b), transp=TRUE)
hasn't been given yet. This uses the triangular structure, while keeping to the original question. This should be faster and more accurate for larger matrices. It might also be worth noting that L <- t(chol(A)) since chol returns an upper triangular matrix.
A <- matrix(c(1,1,1,1,5,5,1,5,14), nrow=3)
# Cholesky decomposition A = LL'
L <- t(chol(A))
# Make some b with known x
x <- c(1, 2, 3)
b <- A %*% x
# Solve
forwardsolve(L, forwardsolve(L, b), transp=TRUE)
Giving the answer:
> forwardsolve(L, forwardsolve(L, b), transp=TRUE)
[,1]
[1,] 1
[2,] 2
[3,] 3
Related
When making predictions for a linear statistical model we usually have a model matrix X of predictors corresponding to the points at which we want to make predictions; a vector of coefficients beta; and a variance-covariance matrix V. Computing the predictions is just X %*% beta. The most straightforward way to compute the variances of the predictions is
diag(X %*% V %*% t(X))
or slightly more efficiently
diag(X %*% tcrossprod(V,X))
However, this is very inefficient, because it constructs an n*n matrix when all we really want is the diagonal. I know I could write some Rcpp-loopy thing that would compute just the diagonal terms, but I'm wondering if there is an existing linear algebra trick in R that will nicely do what I want ... (if someone wants to write the Rcpp-loopy thing for me as an answer I wouldn't object, but I'd prefer a pure-R solution)
FWIW predict.lm seems to do something clever by multiplying X by the inverse of the R component of the QR-decomposition of the lm; I'm not sure that's always going to be available, but it might be a good starting point (see here)
Along the lines of this Octave/Matlab question, for two matrices A and B, we can use the use the fact that the nth diagonal entry of AB will be the product of the nth row of A with the nth column of B. We can naively extend that to the case of three matrices, ABC. I have not considered how to optimize in the case where C=A^T, but aside from that, this code looks like promising speedup:
start_time <- Sys.time()
A=matrix(1:1000000, nrow = 1000, ncol = 1000)
B=matrix(1000000:1, nrow = 1000, ncol = 1000)
# Try one of these two
res=diag(A %*% B %*% t(A)) # ~0.47s
res=rowSums(A * t(B %*% t(A))) # ~0.27s
end_time <- Sys.time()
print(end_time - start_time)
Using tcrossprod did not appear to accelerate the results when I ran this code. However, just using the row-sum-dot-product approach appears to be a lot more efficient already, at least on this silly example, which suggests (though I'm not sure) that rowSums is not computing the full intermediate matrices before returning the diagonal entries, as I'd expect happens with diag.
I am not quite sure how efficient this is,
Find U such that V = U %*% t(U); this is possible since V is cov matrix.
XU = X %*% U
result = apply(XU, 1, function(x) sum(x^2))
Demo
V <- cov(iris[, -5])
X <- as.matrix(iris[1:5, -5])
Using SVD
svd_v <- svd(V)
U <- svd_v$u %*% diag(sqrt(svd_v$d))
XU = X %*% U
apply(XU, 1, function(x) sum(x^2))
# 1 2 3 4 5
#41.35342 39.36286 35.42369 38.25584 40.30839
Another approach - this isn't also going to be faster than #davewy's
U <- chol(V)
XU = (X %*% U)^2
rowSums(XU)
I recently found emulator::quad.diag(), which is just
colSums(crossprod(M, Conj(x)) * x)
This is slightly better than #davewy's solution (although the overall differences are less than I thought they would be anyway).
library(microbenchmark)
microbenchmark(full=diag(A %*% B %*% t(A)),
davewy=rowSums(A * t(B %*% t(A))),
emu = quad.diag(A,B))
Unit: milliseconds
expr min lq mean median uq max neval cld
full 32.76241 35.49665 39.51683 37.63958 41.46561 57.41370 100 c
davewy 22.74787 25.06874 28.42179 26.97330 29.68895 45.38188 100 b
emu 17.68390 20.21322 23.59981 22.09324 24.80734 43.60953 100 a
I have a equation system and I want to solve it using numerical methods. I want to get a close solution given a starting seed. Let me explain.
I have a vector of constants ,X, of values:
X <- (c(1,-2,3,4))
and a vector W of weights:
W <- (c(0.25,0.25,0.25,0.25))
I want that the sum of the components of W will be (sum(W)=1), and the sum of the multiplication of X and W element by element will be a given number N (sum(W*X)=N).
Is there a easy way to do this in R? I have it in Excel, using Solver, but I need to automatize it.
Here is your constant and your target value:
x <- c(1, -2, 3, 4)
n <- 10
You need a function to minimize. The first line contains each of your conditions, and the second line provides a measure of how to combine the errors into a single score. You may want to change the second line. For example, you could make one error term be more heavily weighted than the other using sum(c(1, 5) * errs ^ 2).
fn <- function(w)
{
errs <- c(sum(w) - 1, sum(x * w) - n)
sum(errs ^ 2)
}
The simplest thing is to start with all the weights the same value.
init_w <- rep.int(1 / length(x), length(x))
Use optim to optimize.
optim(init_w, fn)
## $par
## [1] 0.1204827 -1.2438883 1.1023338 1.0212406
##
## $value
## [1] 7.807847e-08
##
## $counts
## function gradient
## 111 NA
##
## $convergence
## [1] 0
##
## $message
## NULL
The par element contains your weights.
There is no unique solution for this problem. If you try other initial values for w you will most likely get different results from optim.
The problem can be formulated as solving an underdetermined system of linear equations.
A <- matrix(c(rep(1,4),x), nrow=2,byrow=TRUE)
b <- matrix(c(1,n), nrow=2)
We seek a solution that satisfies A %*% w = b but which one? Minimum norm solution? Or maybe some other one? There are infinitely many solutions. Solutions can be given using the pseudo-inverse of the matrix A. Use package MASS for this.
library(MASS)
Ag <- ginv(A)
The minimum norm solution is
wmnorm <- Ag %*% b
And check with A %*% wmnorm - b and fn(wmnorm).
See the Wikipedia page System of linear equations
the section Matrix solutions.
The solutions are given by
Az <- diag(nrow=nrow(Ag)) - Ag %*% A
w <- wmnorm + Az %*% z
where z is an arbitrary vector of ncol(Az) elements.
And now generate some solutions and check
xb <- wmnorm
z <- runif(4)
wsol.2 <- xb + Az %*% z
wsol.2
A %*% wsol.2 - b
fn(wsol.2)
z <- runif(4)
wsol.3 <- xb + Az %*% z
wsol.3
A %*% wsol.2 - b
fn(wsol.3)
And you'll see that these two solutions are valid solutions when given as argument to fn. And are quite different from the solution found by optim. You could test this by choosing a different starting point init_w for example by init_w1 <- runif(4)/4.
y <- matrix(c(7, 9, -5, 0, 2, 6), ncol = 1)
try <- t(y)
tryy <- try %*% y
i <- solve(tryy)
h <- y %*% i %*% try
uniroot(as.vector(solve(((1-x) * diag(6)) + h)), c(-Inf, Inf))
Error in (1 - x) * diag(6) : non-conformable arrays
The purpose of this command uniroot(as.vector(solve(((1-x) * diag(6)) + h)), c(-Inf, Inf)) is to solve the characteristics equation det[(1-λ)I+h] = 0
where, λ=eigenvalues , I=identity matrix , h=hat matrix=y(y'y)^(-1)y'
here λ is unknown ,we have to solve for it.
I am not understanding where is the problem here? I have tried as:
as.vector(solve(6*diag(6)+h))
This is not non-conformable. But why is not working inside the uniroot function?
Your question is a bit confusing, so I have to make a couple of assumptions. If you want the eigenvalues of h, then the characteristic equation is:
det(h - I*λ) = 0
not
det[(1-λ)I+h] = 0
So I used the former.
Given the above, the short answer is: do it this way.
f <- function(lambda) det(h -lambda*diag(6))
F <- Vectorize(f)
library(rootSolve)
uniroot.all(F,c(-1000,1000),n=2000)
# [1] 0 1
# or, much more simply
eigen(h)$values
# [1] 1.000000e+00 2.220446e-16 0.000000e+00 -2.731318e-18 -6.876381e-18 -7.365903e-17
So h has 2 eigenvalues, 0 and 1. Note that the built-in function eigen(...) finds 6 roots, but 5 of them are within the machine tolerance of 0.
The question about why your code fails is a bit more involved.
First, your code:
tryy <- try %*% y
is the dot product of y with itself (so, a scalar), returned as a matrix with one element. When you "invert" that using solve(...)
i <- solve(tryy)
you simply take the reciprocal, so i is also a matrix with 1 element. I'm not sure if this is what you had in mind.
Second, uniroot(...) does not work this way. The first argument must be a function; you've passed an expression which depends on x, which in turn is undefined. You could try:
f <- function(x) det(h-x*diag(6))
uniroot(f,c(-Inf,Inf))
but this wouldn't work either because (a) uniroot(...) works on a finite interval, (b) it requires that the function f(...) have different sign at the ends of the interval, and (c) in any event it would return only one root (the smaller one).
So you could use uniroot.all(...) in package rootSolve. uniroot.all(...) also requires a function as it's first argument, but there's a twist: the function must be "vectorized". This means that if you pass a vector of lambda values, f(...) should return a vector of the same length. Fortunately in R there is an easy way to "vectorize" a given function, as in:
F <- Vectorize(f).
Even this has it's limits. uniroot.all(...) also requires a finite interval, so we have to guess what that is, and also it evaluates F on n sub-intervals. So if your interval does not contain all the roots, or if the sub-intervals are not small enough, you will not find all the roots.
Using the built-in eigen(...) function is definitely the best option.
I am hoping to create 3 (non-negative) quasi-random numbers that sum to one, and repeat over and over.
Basically I am trying to partition something into three random parts over many trials.
While I am aware of
a = runif(3,0,1)
I was thinking that I could use 1-a as the max in the next runif, but it seems messy.
But these of course don't sum to one. Any thoughts, oh wise stackoverflow-ers?
This question involves subtler issues than might be at first apparent. After looking at the following, you may want to think carefully about the process that you are using these numbers to represent:
## My initial idea (and commenter Anders Gustafsson's):
## Sample 3 random numbers from [0,1], sum them, and normalize
jobFun <- function(n) {
m <- matrix(runif(3*n,0,1), ncol=3)
m<- sweep(m, 1, rowSums(m), FUN="/")
m
}
## Andrie's solution. Sample 1 number from [0,1], then break upper
## interval in two. (aka "Broken stick" distribution).
andFun <- function(n){
x1 <- runif(n)
x2 <- runif(n)*(1-x1)
matrix(c(x1, x2, 1-(x1+x2)), ncol=3)
}
## ddzialak's solution (vectorized by me)
ddzFun <- function(n) {
a <- runif(n, 0, 1)
b <- runif(n, 0, 1)
rand1 = pmin(a, b)
rand2 = abs(a - b)
rand3 = 1 - pmax(a, b)
cbind(rand1, rand2, rand3)
}
## Simulate 10k triplets using each of the functions above
JOB <- jobFun(10000)
AND <- andFun(10000)
DDZ <- ddzFun(10000)
## Plot the distributions of values
par(mfcol=c(2,2))
hist(JOB, main="JOB")
hist(AND, main="AND")
hist(DDZ, main="DDZ")
just random 2 digits from (0, 1) and if assume its a and b then you got:
rand1 = min(a, b)
rand2 = abs(a - b)
rand3 = 1 - max(a, b)
When you want to randomly generate numbers that add to 1 (or some other value) then you should look at the Dirichlet Distribution.
There is an rdirichlet function in the gtools package and running RSiteSearch('Dirichlet') brings up quite a few hits that could easily lead you to tools for doing this (and it is not hard to code by hand either for simple Dirichlet distributions).
I guess it depends on what distribution you want on the numbers, but here is one way:
diff(c(0, sort(runif(2)), 1))
Use replicate to get as many sets as you want:
> x <- replicate(5, diff(c(0, sort(runif(2)), 1)))
> x
[,1] [,2] [,3] [,4] [,5]
[1,] 0.66855903 0.01338052 0.3722026 0.4299087 0.67537181
[2,] 0.32130979 0.69666871 0.2670380 0.3359640 0.25860581
[3,] 0.01013117 0.28995078 0.3607594 0.2341273 0.06602238
> colSums(x)
[1] 1 1 1 1 1
I would simply randomly select 3 numbers from uniform distribution and then divide by their sum:
n <- 3
x <- runif(n, 0, 1)
y <- x / sum(x)
sum(y) == 1
n could be any number you like.
This problem and the different solutions proposed intrigued me. I did a little test of the three basic algorithms suggested and what average values they would yield for the numbers generated.
choose_one_and_divide_rest
means: [ 0.49999212 0.24982403 0.25018384]
standard deviations: [ 0.28849948 0.22032758 0.22049302]
time needed to fill array of size 1000000 was 26.874945879 seconds
choose_two_points_and_use_intervals
means: [ 0.33301421 0.33392816 0.33305763]
standard deviations: [ 0.23565652 0.23579615 0.23554689]
time needed to fill array of size 1000000 was 28.8600130081 seconds
choose_three_and_normalize
means: [ 0.33334531 0.33336692 0.33328777]
standard deviations: [ 0.17964206 0.17974085 0.17968462]
time needed to fill array of size 1000000 was 27.4301018715 seconds
The time measurements are to be taken with a grain of salt as they might be more influenced by the Python memory management than by the algorithm itself. I'm too lazy to do it properly with timeit. I did this on 1GHz Atom so that explains why it took so long.
Anyway, choose_one_and_divide_rest is the algorithm suggested by Andrie and the poster of the question him/herself (AND): you choose one value a in [0,1], then one in [a,1] and then you look what you have left. It adds up to one but that's about it, the first division is twice as large as the other two. One might have guessed as much ...
choose_two_points_and_use_intervals is the accepted answer by ddzialak (DDZ). It takes two points in the interval [0,1] and uses the size of the three sub-intervals created by these points as the three numbers. Works like a charm and the means are all 1/3.
choose_three_and_normalize is the solution by Anders Gustafsson and Josh O'Brien (JOB). It just generates three numbers in [0,1] and normalizes them back to a sum of 1. Works just as well and surprisingly a little bit faster in my Python implementation. The variance is a bit lower than for the second solution.
There you have it. No idea to what beta distribution these solutions correspond or which set of parameters in the corresponding paper I referred to in a comment but maybe someone else can figure that out.
The simplest solution is the Wakefield package probs() function
probs(3) will yield a vector of three values with a sum of 1
given that you can rep(probs(3),x) where x is "over and over"
no drama
I have a matrix and I would like to know if it is diagonalizable. How do I do this in the R programming language?
If you have a given matrix, m, then one way is the take the eigen vectors times the diagonal of the eigen values times the inverse of the original matrix. That should give us back the original matrix. In R that looks like:
m <- matrix( c(1:16), nrow = 4)
p <- eigen(m)$vectors
d <- diag(eigen(m)$values)
p %*% d %*% solve(p)
m
so in that example p %*% d %*% solve(p) should be the same as m
You can implement the full algorithm to check if the matrix reduces to a Jordan form or a diagonal one (see e.g., this document). Or you can take the quick and dirty way: for an n-dimensional square matrix, use eigen(M)$values and check that they are n distinct values. For random matrices, this always suffices: degeneracy has prob.0.
P.S.: based on a simple observation by JD Long below, I recalled that a necessary and sufficient condition for diagonalizability is that the eigenvectors span the original space. To check this, just see that eigenvector matrix has full rank (no zero eigenvalue). So here is the code:
diagflag = function(m,tol=1e-10){
x = eigen(m)$vectors
y = min(abs(eigen(x)$values))
return(y>tol)
}
# nondiagonalizable matrix
m1 = matrix(c(1,1,0,1),nrow=2)
# diagonalizable matrix
m2 = matrix(c(-1,1,0,1),nrow=2)
> m1
[,1] [,2]
[1,] 1 0
[2,] 1 1
> diagflag(m1)
[1] FALSE
> m2
[,1] [,2]
[1,] -1 0
[2,] 1 1
> diagflag(m2)
[1] TRUE
You might want to check out this page for some basic discussion and code. You'll need to search for "diagonalized" which is where the relevant portion begins.
All symmetric matrices across the diagonal are diagonalizable by orthogonal matrices. In fact if you want diagonalizability only by orthogonal matrix conjugation, i.e. D= P AP' where P' just stands for transpose then symmetry across the diagonal, i.e. A_{ij}=A_{ji}, is exactly equivalent to diagonalizability.
If the matrix is not symmetric, then diagonalizability means not D= PAP' but merely D=PAP^{-1} and we do not necessarily have P'=P^{-1} which is the condition of orthogonality.
you need to do something more substantial and there is probably a better way but you could just compute the eigenvectors and check rank equal to total dimension.
See this discussion for a more detailed explanation.