I'm simulating discrete character data using the function rTraitDisc {ape} in R using a variety of model matrices. I've not encountered any issues with scaling when all state changes are possible. However when I supply an ordered model with 8 or more possible states, the function breaks down and returns the following error:
## library
library(ape)
## read in tree
data("bird.orders")
## build model
model.matrix <- matrix(c(0,0.1,0,0,0,0,0,0,
0.1,0,0.1,0,0,0,0,0,
0,0.1,0,0.1,0,0,0,0,
0,0,0.1,0,0.1,0,0,0,
0,0,0,0.1,0,0.1,0,0,
0,0,0,0,0.1,0,0.1,0,
0,0,0,0,0,0.1,0,0.1,
0,0,0,0,0,0,0.1,0), 8)
## run function
rTraitDisc(phy = bird.orders, model = model.matrix)
Error message:
Error in sample.int(k, size = 1, FALSE, prob = p) : negative probability
Having dug a little deeper, it seems that when there are 8 or more states but only one possible transition (e.g. if the ancestral state is 0, only a transition to state 1 should be possible in an ordered matrix), the function matexpo produces a probability matrix with negative values for the shortest branch of the tree (0.5). As these probabilities are used by sample.int as the "prob" argument, the negative probabilities cause the function to break down.
## get number of states
k <- ncol(model.matrix)
## get equilibrium relative frequencies
freq = rep(1/k, k)
## match number of elements in model
freq <- rep(freq, each = k)
## get Q matrix
Q <- model.matrix * freq
diag(Q) <- 0
diag(Q) <- -rowSums(Q)
## get minimum edge length
min.el <- min(bird.orders$edge.length)
## run matexpo
matexpo(Q*min.el)
How do I deal with these negative values in this context? Is there a correction I can/should apply?
Related
I am attempting to run a two state HMM using a lognormal distribution. I have read Michelot and Langrock (2019) regarding choosing starting parameters through inspecting the data in a histogram and then running iterations in parallel, which has worked for my gamma distribution. Identifying the starting parameters for the lognormal distribution is troubling me however. Do I plot the log of my step length distribution then attempt extracting starting parameters or use the same starting parameters as my gamma distribution and rely on stepDist="lnorm"?
My code for the lognormal attempt currently looks like this:
ncores <- detectCores() - 1
cl <- makeCluster(getOption("cl.cores", ncores))
clusterExport(cl, list("data", "fitHMM"))
niter <- 20
allPar0 <- lapply(as.list(1:niter), function(x) {
stepMean0 <- runif(2,
min = c(x,y),
max = c(y,z))
stepSD0 <- runif(2,
min = c(x,y),
max = c(y,z))
angleMean0 <- c(0, 0)
angleCon0 <- runif(2,
min = c(a,b),
max = c(a,b))
stepPar0 <- c(stepMean0, stepSD0)
anglePar0 <- c(angleMean0, angleCon0)
return(list(step = stepPar0, angle = anglePar0))
})
# Fit the niter models in parallel
logP <- parLapply(cl = cl, X = allPar0, fun = function(par0) {
m <- fitHMM(data = data, nbStates = 2, stepDist = "lnorm", stepPar0 = par0$step,
anglePar0 = par0$angle)
return(m)
})
# Extract likelihoods of fitted models
likelihoodL <- unlist(lapply(logP, function(m) m$mod$minimum))
likelihoodL
# Index of best fitting model (smallest negative log-likelihood)
whichbestpL <- which.min(likelihoodL)
bestL <- logP[[whichbestpL]]
bestL
If I use negative values from plotting the log of the step length of the data then I get the error:
Error in checkForRemoteErrors(val) :
7 nodes produced errors; first error: Check the step parameters bounds (the initial parameters should be strictly between the bounds of their parameter space).
If I use the same starting parameter values that I used for my gamma distribution then I get the error
Error in unserialize(node$con) :
embedded nul in string: 'X\n\0\0\0\003\0\004\002\0\0\003\005\0\0\0'
Please could someone shed some light on how I'm failing at this?
Thank you!
Unfortunately, I can't tell for sure what the problem is from the code you included. If you don't get an error when you run fitHMM outside of parLapply, then it suggests that the problem is in how you choose the values of x, y, and z in your code.
The first parameter of the log-normal distribution can be negative or positive, and it is actually the mean of the logarithm of the step length. So, to find good starting values for this, you should look at a histogram of the log step lengths (e.g., following the dedicated moveHMM vignette). The second parameter is the standard deviation of the log step lengths, and this should be strictly positive (but could also be chosen based on the spread of the histogram of log step lengths).
To summarise, you should choose all the initial values based on plots of the log step lengths (rather than the step lengths themselves), and you should not use the same ranges of values for stepMean0 and stepSD0 (because the former can be negative or positive, whereas the latter is positive). Hopefully, this should help you choose x, y, and z.
I'm doing a beta regression in R, which requires values between 0 and 1, endpoints excluded, i.e. (0,1) instead of [0,1].
I have some 0 and 1 values in my dataset, so I'd like to convert them to the smallest possible neighbor, such as 0.0000...0001 and 0.9999...9999. I've used .Machine$double.xmin (which gives me 2.225074e-308), but betareg() still gives an error:
invalid dependent variable, all observations must be in (0, 1)
If I use 0.000001 and 0.999999, I got a different set of errors:
1: In betareg.fit(X, Y, Z, weights, offset, link, link.phi, type, control) :
failed to invert the information matrix: iteration stopped prematurely
2: In sqrt(wpp) :
Error in chol.default(K) :
the leading minor of order 4 is not positive definite
Only if I use 0.0001 and 0.9999 I can run without errors. Is there any way I can improve this minimum values with betareg? Or should I just be happy with that?
Try it with eps (displacement from 0 and 1) first equal to 1e-4 (as you have here) and then with 1e-3. If the results of the models don't differ in any way you care about, that's great. If they are, you need to be very careful, because it suggests your answers will be very sensitive to assumptions.
In the example below the dispersion parameter phi changes a lot, but the intercept and slope parameter don't change very much.
If you do find that the parameters change by a worrying amount for your particular data, then you need to think harder about the process by which zeros and ones arise, and model that process appropriately, e.g.
a censored-data model: zero/one arise through a minimum/maximum detection threshold, models the zero/one values as actually being somewhere in the tails or
a hurdle/zero-one inflation model: zeros and ones arise through a separate process from the rest of the data, use a binomial or multinomial model to characterize zero vs. (0,1) vs. one, then use a Beta regression on the (0,1) component)
Questions about these steps are probably more appropriate for CrossValidated than for SO.
sample data
set.seed(101)
library(betareg)
dd <- data.frame(x=rnorm(500))
rbeta2 <- function(n, prob=0.5, d=1) {
rbeta(n, shape1=prob*d, shape2=(1-prob)*d)
}
dd$y <- rbeta2(500,plogis(1+5*dd$x),d=1)
dd$y[dd$y<1e-8] <- 0
trial fitting function
ss <- function(eps) {
dd <- transform(dd,
y=pmin(1-eps,pmax(eps,y)))
m <- try(betareg(y~x,data=dd))
if (inherits(m,"try-error")) return(rep(NA,3))
return(coef(m))
}
ss(0) ## fails
ss(1e-8) ## fails
ss(1e-4)
## (Intercept) x (phi)
## 0.3140810 1.5724049 0.7604656
ss(1e-3) ## also fails
ss(1e-2)
## (Intercept) x (phi)
## 0.2847142 1.4383922 1.3970437
ss(5e-3)
## (Intercept) x (phi)
## 0.2870852 1.4546247 1.2029984
try it for a range of values
evec <- seq(-4,-1,length=51)
res <- t(sapply(evec, function(e) ss(10^e)) )
library(ggplot2)
ggplot(data.frame(e=10^evec,reshape2::melt(res)),
aes(e,value,colour=Var2))+
geom_line()+scale_x_log10()
I have a two-dimensional point pattern (no marks) and I am trying to test for clustering in the presence of spatial inhomogeneity using envelopes and the inhomogenous pair correlation function. I am estimating an inhomogenous intensity function for the data using the density.ppp function. Here is some sample data:
x y
1 533.03 411.58
2 468.39 622.92
3 402.86 530.94
4 427.13 616.81
5 495.20 680.62
6 566.61 598.99
7 799.03 585.16
8 1060.09 544.23
9 144.66 747.40
10 138.14 752.92
11 449.49 839.15
12 756.45 713.72
13 741.01 728.41
14 760.22 740.28
15 802.34 756.21
16 799.04 764.89
17 773.81 771.97
18 768.41 720.07
19 746.14 754.11
20 815.40 765.14
There are ~1700 data points overall
Here is my code:
library("spatstat")
WT <- read.csv("Test.csv")
colnames(WT) <- c("x","y")
#determine bounding window
win <- ripras(WT)
unitname(win) <- c("micrometer")
#convert to ppp data class
WT.ppp <- as.ppp(WT, win)
plot(WT.ppp)
#estimate intensity function using cross validation
I <- density.ppp(WT.ppp,sigma=bw.diggle(WT.ppp),adjust=0.3,kernal="epanechnikov")
plot(I)
#predetermined r values for PCF
radius <- seq(from = 0, to = 50, by = 0.5)
#use envelopes to test the null hypothesis (ie. inhomogenous poisson process)
PCF_envelopes <- envelope(WT.ppp,divisor="d", pcfinhom,r = radius,nsim=10,simulate=expression(rpoispp(I)) )
When I run rpoisspp(I), I get the following error:
Error in sample.int(npix, size = ni, replace = TRUE, prob = lpix) :
negative probability
I can't seem to figure out what the issue is....any suggestions?
Thanks for your help!
This is happening because the image I contains some negative values, probably very small values but negative. You can check that by computing range(I) or min(I) or any(I < 0).
The help for density.ppp says that the result may contain negative values (very small ones) due to numerical error. To remove these, you need to set positive=TRUE in the call to density.ppp.
By the way, the argument kernel has been mis-spelt in the code above. Also the vector r is too coarsely spaced - you would be better to leave this argument un-specified. Also you don't need to type density.ppp, just density.
I tried to use princomp() and principal() to do PCA in R with data set USArressts. However, I got two different results for loadings/rotaion and scores.
First, I centered and normalised the original data frame so it is easier to compare the outputs.
library(psych)
trans_func <- function(x){
x <- (x-mean(x))/sd(x)
return(x)
}
A <- USArrests
USArrests <- apply(USArrests, 2, trans_func)
princompPCA <- princomp(USArrests, cor = TRUE)
principalPCA <- principal(USArrests, nfactors=4 , scores=TRUE, rotate = "none",scale=TRUE)
Then I got the results for the loadings and scores using the following commands:
princompPCA$loadings
principalPCA$loadings
Could you please help me to explain why there is a difference? and how can we interprete these results?
At the very end of the help document of ?principal:
"The eigen vectors are rescaled by the sqrt of the eigen values to produce the component loadings more typical in factor analysis."
So principal returns the scaled loadings. In fact, principal produces a factor model estimated by the principal component method.
In 4 years, I would like to provide a more accurate answer to this question. I use iris data as an example.
data = iris[, 1:4]
First, do PCA by the eigen-decomposition
eigen_res = eigen(cov(data))
l = eigen_res$values
q = eigen_res$vectors
Then the eigenvector corresponding to the largest eigenvalue is the factor loadings
q[,1]
We can treat this as a reference or the correct answer. Now we check the results by different r functions.
First, by function 'princomp'
res1 = princomp(data)
res1$loadings[,1]
# compare with
q[,1]
No problem, this function actually just return the same results as 'eigen'. Now move to 'principal'
library(psych)
res2 = principal(data, nfactors=4, rotate="none")
# the loadings of the first PC is
res2$loadings[,1]
# compare it with the results by eigendecomposition
sqrt(l[1])*q[,1] # re-scale the eigen vector by sqrt of eigen value
You may find they are still different. The problem is the 'principal' function does eigendecomposition on the correlation matrix by default. Note: PCA is not invariant with rescaling the variables. If you modify the code as
res2 = principal(data, nfactors=4, rotate="none", cor="cov")
# the loadings of the first PC is
res2$loadings[,1]
# compare it with the results by eigendecomposition
sqrt(l[1])*q[,1] # re-scale the eigen vector by sqrt of eigen value
Now, you will get the same results as 'eigen' and 'princomp'.
Summarize:
If you want to do PCA, you'd better apply 'princomp' function.
PCA is a special case of the Factor model or a simplified version of the factor model. It is just equivalent to eigendecomposition.
We can apply PCA to get an approximation of a factor model. It doesn't care about the specific factors, i.e. epsilons in a factor model. So, if you change the number of factors in your model, you will get the same estimations of the loadings. It is different from the maximum likelihood estimation.
If you are estimating a factor model, you'd better use 'principal' function, since it provides more functions, like rotation, calculating the scores by different methods, and so on.
Rescale the loadings of a PCA model doesn't affect the results too much. Since you still project the data onto the same optimal direction, i.e. maximize the variation in the resulting PC.
ev <- eigen(R) # R is a correlation matrix of DATA
ev$vectors %*% diag(ev$values) %*% t(ev$vectors)
pc <- princomp(scale(DATA, center = F, scale = T),cor=TRUE)
p <-principal(DATA, rotate="none")
#eigen values
ev$values^0.5
pc$sdev
p$values^0.5
#eigen vectors - loadings
ev$vectors
pc$loadings
p$weights %*% diag(p$values^0.5)
pc$loading %*% diag(pc$sdev)
p$loadings
#weights
ee <- diag(0,2)
for (j in 1:2) {
for (i in 1:2) {
ee[i,j] <- ev$vectors[i,j]/p$values[j]^0.5
}
};ee
#scores
s <- as.matrix(scale(DATA, center = T, scale = T)) %*% ev$vectors
scale(s)
p$scores
scale(pc$scores)
I have to create a model which is a mixture of a normal and log-normal distribution. To create it, I need to estimate the 2 covariance matrixes and the mixing parameter (total =7 parameters) by maximizing the log-likelihood function. This maximization has to be performed by the nlm routine.
As I use relative data, the means are known and equal to 1.
I’ve already tried to do it in 1 dimension (with 1 set of relative data) and it works well. However, when I introduce the 2nd set of relative data I get illogical results for the correlation and a lot of warnings messages (at all 25).
To estimate these parameters I defined first the log-likelihood function with the 2 commands dmvnorm and dlnorm.plus. Then I assign starting values of the parameters and finally I use the nlm routine to estimate the parameters (see script below).
`P <- read.ascii.grid("d:/Documents/JOINT_FREQUENCY/grid_E727_P-3000.asc", return.header=
FALSE );
V <- read.ascii.grid("d:/Documents/JOINT_FREQUENCY/grid_E727_V-3000.asc", return.header=
FALSE );
p <- c(P); # tranform matrix into a vector
v <- c(V);
p<- p[!is.na(p)] # removing NA values
v<- v[!is.na(v)]
p_rel <- p/mean(p) #Transforming the data to relative values
v_rel <- v/mean(v)
PV <- cbind(p_rel, v_rel) # create a matrix of vectors
L <- function(par,p_rel,v_rel) {
return (-sum(log( (1- par[7])*dmvnorm(PV, mean=c(1,1), sigma= matrix(c(par[1]^2, par[1]*par[2]
*par[3],par[1]*par[2]*par[3], par[2]^2 ),nrow=2, ncol=2))+
par[7]*dlnorm.rplus(PV, meanlog=c(1,1), varlog= matrix(c(par[4]^2,par[4]*par[5]*par[6],par[4]
*par[5]*par[6],par[5]^2), nrow=2,ncol=2)) )))
}
par.start<- c(0.74, 0.66 ,0.40, 1.4, 1.2, 0.4, 0.5) # log-likelihood estimators
result<-nlm(L,par.start,v_rel=v_rel,p_rel=p_rel, hessian=TRUE, iterlim=200, check.analyticals= TRUE)
Messages d'avis :
1: In log(eigen(sigma, symmetric = TRUE, only.values = TRUE)$values) :
production de NaN
2: In sqrt(2 * pi * det(varlog)) : production de NaN
3: In nlm(L, par.start, p_rel = p_rel, v_rel = v_rel, hessian = TRUE) :
NA/Inf replaced by maximum positive value
4: In log(eigen(sigma, symmetric = TRUE, only.values = TRUE)$values) :
production de NaN
…. Until 25.
par.hat <- result$estimate
cat("sigN_p =", par[1],"\n","sigN_v =", par[2],"\n","rhoN =", par[3],"\n","sigLN_p =", par [4],"\n","sigLN_v =", par[5],"\n","rhoLN =", par[6],"\n","mixing parameter =", par[7],"\n")
sigN_p = 0.5403361
sigN_v = 0.6667375
rhoN = 0.6260181
sigLN_p = 1.705626
sigLN_v = 1.592832
rhoLN = 0.9735974
mixing parameter = 0.8113369`
Does someone know what is wrong in my model or how should I do to find these parameters in 2 dimensions?
Thank you very much for taking time to look at my questions.
Regards,
Gladys Hertzog
When I do these kind of optimization problems, I find that it's important to make sure that all the variables that I'm optimizing over are constrained to plausible values. For example, standard deviation variables have to be positive, and from knowledge of the situation that I'm modelling I'll probably be able to put an upper bound all my standard deviation variables as well. So if s is one of my standard deviation variables, and if m is the maximum value that I want it to take, instead of working with s I'll solve for the variable z which is related to s via
s = m/(1+e-z)
In that formula, z is unconstrained, but s must lie between 0 and m. This is vital because optimization routines where the variables are not constrained to take plausible values will often try completely implausible values while they're trying to bound the solution. Implausible values often cause problems with e.g. precision, that then results in NaN's etc. The general formula that I use for constraining a single variable x to lie between a and b is
x = a + (b - a)/(1+e-z)
However, regarding your particular problem where you're looking for covariance matrices, a more sophisticated approach is necessary than simply bounding all the individual variables. Covariance matrices must be positive semi-definite, so if you're simply optimizing the individual values in the matrix, the optimization will probably fail (producing NaN's) if a matrix which isn't positive definite is fed into the likelihood function. To get round this problem, one approach is to solve for the Cholesky decomposition of the covariance matrix instead of the covariance matrix itself. My guess is that this is probably what's causing your optimization to fail.