rstan Install issues

[jvolpe...@gmail.com]

I'm having problems installing for the first time rstan. I'm using R 3.3.2 and followed the RStan Getting Started document. I'm getting the following errors when trying the example "rats"

> y <- read.table('rats.txt', header = TRUE)

> x <- c(8, 15, 22, 29, 36)

> xbar <- mean(x)

> N <- nrow(y)

> T <- ncol(y)

> rats_fit <- stan(file = 'rats.stan')

DIAGNOSTIC(S) FROM PARSER:

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

In file included from C:/Program Files/R/R-3.3.2/library/BH/include/boost/config.hpp:39:0,

                 from C:/Program Files/R/R-3.3.2/library/BH/include/boost/math/tools/config.hpp:13,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core/var.hpp:7,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core/gevv_vvv_vari.hpp:5,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core.hpp:12,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/mat.hpp:4,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math.hpp:4,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/src/stan/model/model_header.hpp:4,

                 from file1570189b551b.cpp:8:

C:/Program Files/R/R-3.3.2/library/BH/include/boost/config/compiler/gcc.hpp:186:0: warning: "BOOST_NO_CXX11_RVALUE_REFERENCES" redefined

 #  define BOOST_NO_CXX11_RVALUE_REFERENCES

 ^

<command-line>:0:0: note: this is the location of the previous definition

cc1plus.exe: warning: unrecognized command line option "-Wno-ignored-attributes"

Error in tryCatch(evalq(getNamespace("Rcpp"), <environment>), error = function (x)  : 

  unused arguments (error = function (x) 

x, interrupt = function (x) 

x)

In addition: Warning messages:

1: In is.na(x) : is.na() applied to non-(list or vector) of type 'NULL'

2: In FUN(X[[i]], ...) : data with name y is not numeric and not used

[Ben Goodrich]

On Thursday, January 12, 2017 at 11:52:05 AM UTC-5, jvolpe...@gmail.com wrote:

I'm having problems installing for the first time rstan. I'm using R 3.3.2 and followed the RStan Getting Started document. I'm getting the following errors when trying the example "rats"

It may work if you add an as.matrix() to
 

> y <- read.table('rats.txt', header = TRUE)

so that it is

y <- as.matrix(read.table('rats.txt', header = TRUE))

I have adjusted that line in the wiki. If not, then you have some sort of an installation problem.

Ben

[Atul Ingle]

It works for me when replace all the "<-" with "=" in rats.rstan and typecast y to be a matrix
y = as.matrix(y)
before calling stan().

On Thursday, January 12, 2017 at 11:52:05 AM UTC-5, jvolpe...@gmail.com wrote:

[Bob Carpenter]

We'll take a pass through all the models soon and clean this up.
The warning isn't fatal---you can still use <- instead of =, but
we'll eliminate <- at some point in the future.

- Bob

> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[jvolpe...@gmail.com]

Same problem with the as.matrix added.

> y <- as.matrix(read.table('rats.txt', header = TRUE))

> x <- c(8, 15, 22, 29, 36)

> xbar <- mean(x)

> N <- nrow(y)

> T <- ncol(y)

> rats_fit <- stan(file = 'rats.stan')

DIAGNOSTIC(S) FROM PARSER:

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

Warning (non-fatal): assignment operator <- deprecated in the Stan language; use = instead.

In file included from C:/Program Files/R/R-3.3.2/library/BH/include/boost/config.hpp:39:0,

                 from C:/Program Files/R/R-3.3.2/library/BH/include/boost/math/tools/config.hpp:13,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core/var.hpp:7,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core/gevv_vvv_vari.hpp:5,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/core.hpp:12,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math/rev/mat.hpp:4,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/stan/math.hpp:4,

                 from C:/Program Files/R/R-3.3.2/library/StanHeaders/include/src/stan/model/model_header.hpp:4,

                 from file101c444cad6.cpp:8:

C:/Program Files/R/R-3.3.2/library/BH/include/boost/config/compiler/gcc.hpp:186:0: warning: "BOOST_NO_CXX11_RVALUE_REFERENCES" redefined

 #  define BOOST_NO_CXX11_RVALUE_REFERENCES

 ^

<command-line>:0:0: note: this is the location of the previous definition

cc1plus.exe: warning: unrecognized command line option "-Wno-ignored-attributes"

Error in tryCatch(evalq(getNamespace("Rcpp"), <environment>), error = function (x)  : 

  unused arguments (error = function (x) 

x, interrupt = function (x) 

x)

> 

[Marina Golivets]

Hello,

I am getting a similar error when passing a list of matrices to Stan:

"is.na() applied to non-(list or vector) of type 'NULL'data with name V_d is not numeric and not usedError : variable does not exist; processing stage=data initialization; variable name=V_d; base type=matrix_d

failed to create the sampler; sampling not done"

The code had worked fine before I reinstalled rstan on my computer. I tried reinstalling rstan once again following all the instructions, but it didn't solve the issue. 

Also, I tried adding -Wno-macro-redefined to CXXFLAGS in .R/Makevars to suppress warnings (as suggested here: https://github.com/stan-dev/rstan/issues/372), but, again, it didn't work.

Marina

четвер, 12 січня 2017 р. 16:15:18 UTC-5 користувач jvolpe...@gmail.com написав:

[Ben Goodrich]

On Sunday, January 29, 2017 at 12:28:24 PM UTC-5, Marina Golivets wrote:

I am getting a similar error when passing a list of matrices to Stan:

This has been fixed on GitHub: https://github.com/stan-dev/rstan/issues/384

Ben

[Marina Golivets]

Thank you, Ben! I am not sure if I can apply it to my case, however, because I was passing a big list (>1000) of variance-covariance matrices. If I pass an array instead and declare it as real[ , , ], I don't see how I can redeclare it as an array of variance-covariance matrices. Here is the code I was using before:

data {

...

matrix[Kmax, Kmax] V_d[K]; // sampling variance-covariance matrices 

}

...

model {

int pos;

    pos = 1;

    for (i in 1:K) {

    segment(d, pos, L[i]) ~ multi_normal(segment(theta, pos, L[i]), block (V_d[i], 1, 1, L[i], L[i]));

    pos = pos + L[i];

    }

}

Thank you for any help,

Marina

неділя, 29 січня 2017 р. 13:01:42 UTC-5 користувач Ben Goodrich написав:

[Ben Goodrich]

On Sunday, January 29, 2017 at 1:41:47 PM UTC-5, Marina Golivets wrote:

Thank you, Ben! I am not sure if I can apply it to my case, however, because I was passing a big list (>1000) of variance-covariance matrices. If I pass an array instead and declare it as real[ , , ], I don't see how I can redeclare it as an array of variance-covariance matrices. Here is the code I was using before:

You can pass an R array that is 1000 by 3 by 3 or whatever to the data block of a Stan program that has

cov_matrix[3] Sigma[1000];

Or you can install the fixed rstan from GitHub via devtools::install_github.

Ben

[Marina Golivets]

I'm afraid that if I pass an R array and declare it as cov_matrix[3] Sigma[1000], the orders of dimensions will not match. It's gonna be [3, 3, 1000] vs [1000, 3, 3].

So I might reinstall the older version of rstan.

Thanks,

Marina

неділя, 29 січня 2017 р. 13:44:32 UTC-5 користувач Ben Goodrich написав:

[Ben Goodrich]

On Sunday, January 29, 2017 at 2:00:37 PM UTC-5, Marina Golivets wrote:

I'm afraid that if I pass an R array and declare it as cov_matrix[3] Sigma[1000], the orders of dimensions will not match. It's gonna be [3, 3, 1000] vs [1000, 3, 3].

The dimensions of Sigma in Stan are 1000 x 3 x 3, just like in R. You don't read from left to right.
 

 So I might reinstall the older version of rstan.

Those versions have a broken NUTS algorithm.

Ben
 

[Marina Golivets]

I might be completely missing something out, but for a three-dimensional R array the dimensions would be listed as [3, 3, 1000]. For example:

> a <- matrix(rep(1, 4), byrow=T, 2, 2)

> b <- matrix(rep(2, 4), byrow=T, 2, 2)

> c <- matrix(rep(3, 4),byrow=T, 2, 2)

> M <-  cbind(a, b, c)

> dim(M) <- c(3, 2, 2)

> M

, , 1

     [,1] [,2]

[1,]    1    1

[2,]    1    2

[3,]    1    2

, , 2

     [,1] [,2]

[1,]    2    3

[2,]    2    3

[3,]    3    3

> dim(M) <- c(2, 2, 3)

> M

, , 1

     [,1] [,2]

[1,]    1    1

[2,]    1    1

, , 2

     [,1] [,2]

[1,]    2    2

[2,]    2    2

, , 3

     [,1] [,2]

[1,]    3    3

[2,]    3    3

Best,

Marina

--
You received this message because you are subscribed to a topic in the Google Groups "Stan users mailing list" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/stan-users/PgT09wl3psQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to stan-users+unsubscribe@googlegroups.com.

To post to this group, send email to stan-...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Best regards, 

Marina Golivets

[Ben Goodrich]

On Sunday, January 29, 2017 at 3:11:31 PM UTC-5, Marina Golivets wrote:

I might be completely missing something out, but for a three-dimensional R array the dimensions would be listed as [3, 3, 1000]. For example:

> a <- matrix(rep(1, 4), byrow=T, 2, 2)

> b <- matrix(rep(2, 4), byrow=T, 2, 2)

> c <- matrix(rep(3, 4),byrow=T, 2, 2)

> M <-  cbind(a, b, c)

> dim(M) <- c(3, 2, 2)

Iff the data block of your Stan program included

matrix[2,2] M[3];

this M would be fine to pass from R.
 

> dim(M) <- c(2, 2, 3)

If your data block of the Stan program included

matrix[2,2] M[3];

this would yield an error because Stan is expecting M to be an R array that is 3x2x2.

Ben

[Marina Golivets]

Thanks, Ben! I rearranged the data in the R array that I was passing to Stan and it worked just fine.

Marina

неділя, 29 січня 2017 р. 18:01:13 UTC-5 користувач Ben Goodrich написав: