[julia-users] Re: Has anyone successfully performed probit or logit regression in Julia?
This works for me:
```
*julia **fit(GeneralizedLinearModel,Y~X,data,Binomial(),ProbitLink())*
*DataFrameRegressionModel{GeneralizedLinearModel,Float64}:*
*Coefficients:*
*Estimate Std.Error z value Pr(|z|)*
*(Intercept) 0.430727 1.980190.217518 0.8278*
*X2.37745e-17 0.91665 2.59362e-17 1.*
*julia **fit(GeneralizedLinearModel,Y~X,data,Binomial(),LogitLink())*
*DataFrameRegressionModel{GeneralizedLinearModel,Float64}:*
*Coefficients:*
* Estimate Std.Error z value Pr(|z|)*
*(Intercept) 0.693147 3.24037 0.21391 0.8306*
*X-7.44332e-17 1.5 -4.96221e-17 1.*
*```*
On Sunday, August 31, 2014 1:27:15 PM UTC-4, Bradley Setzler wrote:
Has anyone successfully performed probit or logit regression in Julia? The
GLM
documentation https://github.com/JuliaStats/GLM.jl does not provide a
generalizable example of how to use glm(). It gives a Poisson example
without any suggestion of how to switch from Poisson to some other type.
*Using the Poisson example from GLM documentation works:*
julia X = [1;2;3.]
julia Y = [1;0;1.]
julia data = DataFrame(X=X,Y=Y)
julia fit(GeneralizedLinearModel, Y ~ X,data, Poisson())
DataFrameRegressionModel{GeneralizedLinearModel,Float64}:
Coefficients:
Estimate Std.Error z value Pr(|z|)
(Intercept) -0.405465 1.87034 -0.216787 0.8284
X -3.91448e-17 0.8658 -4.52123e-17 1.
*But does not generalize:*
julia fit(GeneralizedLinearModel, Y ~ X ,data, Logit())
ERROR: Logit not defined
julia fit(GeneralizedLinearModel, Y ~ X, data, link=:ProbitLink)
ERROR: `fit` has no method matching fit(::Type{GeneralizedLinearModel},
::Array{Float64,2}, ::Array{Float64,1})
julia fit(GeneralizedLinearModel, Y ~ X, data,
family=binomial,link=probit)
ERROR: `fit` has no method matching fit(::Type{GeneralizedLinearModel},
::Array{Float64,2}, ::Array{Float64,1})
and a dozen other similar attempts fail.
Thanks,
Bradley
[julia-users] trouble building julia on mac
I just tried building the latest master on my mac; I get the following error. Any ideas on what to do -- I'm pretty new to this. Googling tells me that this is an issue with clang in Xcode 5.1 but what can I do to fix this? Thanks, Adam adams-air:julia adamkapor$ make Making install in support Making install in kernel libtool: compile: clang -stdlib=libc++ -mmacosx-version-min=10.7 -DHAVE_CONFIG_H -I. -I.. -I../simd -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -MT align.lo -MD -MP -MF .deps/align.Tpo -c align.c -fno-common -DPIC -o .libs/align.o clang: error: unknown argument: '-malign-double' [-Wunused-command-line-argument-hard-error-in-future] clang: note: this will be a hard error (cannot be downgraded to a warning) in the future make[5]: *** [align.lo] Error 1 make[4]: *** [install-recursive] Error 1 make[3]: *** [/Applications/julia/usr/lib/libfftw3f.dylib] Error 2 make[2]: *** [install-fftw] Error 2 make[1]: *** [julia-release] Error 2 make: *** [release] Error 2
Re: [julia-users] trouble building julia on mac
Thank you Elliot and Jameson. I indeed recently upgraded Xcode and clang,
and `make -C deps distclean-fftw` did the trick.
On Wed, Aug 13, 2014 at 2:39 PM, Jameson Nash vtjn...@gmail.com wrote:
Especially true if you just upgraded you OS / compiler, you should start
again with a fresh clone
On Wednesday, August 13, 2014, Elliot Saba staticfl...@gmail.com wrote:
Adam, is this julia codebase a fresh clone, or have you been building out
of this directory for a while? If it's not 100% fresh, you might want to
try running `make -C deps distclean-fftw` and then trying again. It's
possible that we fixed this issue a while back, and the cached configure
state of your fftw is behind a bit and needs to be reconfigured.
-E
On Wed, Aug 13, 2014 at 2:29 PM, Jeff Waller truth...@gmail.com wrote:
Likewise no problems compiling using an older version of clang (Xcode
5.0.2), but I do see that flag
bizarro% clang --version
Apple LLVM version 5.0 (clang-500.2.79) (based on LLVM 3.3svn)
Target: x86_64-apple-darwin13.3.0
Thread model: posix
It looks like it's in there. There is a test to check (see below) but
the compiler doesn't fail, it just warns so the test is perhaps concluding
that yep, supported which is technically true.
bizarro% find . -type f -exec grep malign-double {} /dev/null \;
...
./deps/fftw-3.3.3-double/config.log:clang: warning: argument unused
during compilation: '-malign-double'
./deps/fftw-3.3.3-double/config.log:configure:14242: clang
-stdlib=libc++ -mmacosx-version-min=10.7 -m64 -c -O3 -fomit-frame-pointer
-mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns
-ffast-math
...
./deps/fftw-3.3.3-single/configure:$as_echo_n checking whether C
compiler accepts -malign-double... 6; }
./deps/fftw-3.3.3-single/configure: CFLAGS=-malign-double
./deps/fftw-3.3.3-single/configure: CFLAGS=$CFLAGS -malign-double
./deps/fftw-3.3.3-single/configure: # -malign-double for x86 systems
./deps/fftw-3.3.3-single/configure: { $as_echo
$as_me:${as_lineno-$LINENO}: checking whether C compiler accepts
-malign-double 5
./deps/fftw-3.3.3-single/configure:$as_echo_n checking whether C
compiler accepts -malign-double... 6; }
./deps/fftw-3.3.3-single/configure: CFLAGS=-malign-double
./deps/fftw-3.3.3-single/configure: CFLAGS=$CFLAGS -malign-double
./deps/fftw-3.3.3-single/configure: { $as_echo
$as_me:${as_lineno-$LINENO}: checking whether C compiler accepts
-malign-double 5
./deps/fftw-3.3.3-single/configure:$as_echo_n checking whether C
compiler accepts -malign-double... 6; }
./deps/fftw-3.3.3-single/configure: CFLAGS=-malign-double
./deps/fftw-3.3.3-single/configure: CFLAGS=$CFLAGS -malign-double
...
Binary file ./usr/lib/libfftw3.3.dylib matches
Binary file ./usr/lib/libfftw3.a matches
Binary file ./usr/lib/libfftw3f.3.dylib matches
Binary file ./usr/lib/libfftw3f.a matches
[julia-users] Re: parallel for loop suggestions?
I had a similar issue (former Matlab user, big likelihood computation that needs a giant read-only matrix for each observation.) I haven't dealt with preallocating outputs, but for the read-only matrices can you do something like: mybigmatrix = ... # @everywhere const mybigmatrix = @fetchfrom(1,mybigmatrix) @everywhere function likelihood(params,data_i,mybigmatrix) ... # end @everywhere obj(params,data_i) = likelihood(params,data_i,mybigmatrix) @parallel (+) for ii in observations #or use pmap obj(params,ii) end ``` I didn't know the answer beforehand, so I worked out an example. https://gist.github.com/adamkapor/52ee025b243bf3f5f990 I may be doing this badly but as far as I can tell, you can do fine with @parallel or pmap, but naive @parallel copies everything each call. Adam On Monday, July 7, 2014 2:22:56 PM UTC-4, Thomas Covert wrote: I was indeed planning on using parallel map, but I'm still not clear on whether it will allocate copies of arrays for each worker. On Monday, July 7, 2014 1:21:13 PM UTC-5, Sid wrote: For (1), perhaps something like the parallel map mentioned here ( http://julia.readthedocs.org/en/latest/manual/parallel-computing/) would work? On Monday, July 7, 2014 2:00:43 PM UTC-4, Thomas Covert wrote: Any suggestions about how to do this? Thanks in advance. -Thom On Friday, July 4, 2014 9:15:27 AM UTC-5, Thomas Covert wrote: Hi Julia-Users, I'm trying to parallelize a likelihood calculation in which each step requires read-only access to several large matrices (10,000 by 3000 in one case, 3,000 square in another case, etc). 1) If I do a @parallel for, is Julia going to create copies of each of these matrices? If so, would specifying them as DArray or SharedArray prevent this memory allocation? There are enough of these matrices (and they are big enough) that copying them for parallel programming purposes pretty much defeats the benefits of parallel execution in my setting, I think. I'd really like to avoid that if possible. 2) Each step of the for loop also generates medium sized matrices (with no gradients, its a few 100x1 vectors per loop iteration, with gradients I'd also add some 100x20 matrices). Should I be writing these outputs to a SharedArray or a DArray? I think my head is still stuck in the MATLAB parfor/sliced-array mindset, so any advice from previous MATLAB users would be especially helpful. Thanks. -Thom
[julia-users] typeassert when assigning to elements of array
I'm not sure what :(::) is supposed to do below, but not giving an error at
top level suggests to me that x[1] is an expression computing a value. In
any case, I was surprised that += runs but doesn't `setindex!` x
julia x=zeros(2)
2-element Array{Float64,1}:
0.0
0.0
julia x[1]::Float64 += 1.
1.0
julia x
2-element Array{Float64,1}:
0.0
0.0
julia x[1]::Float64 = 1.
ERROR: syntax: invalid assignment location x[1]
julia myf(x) = x[1]::Float64+=1
myf (generic function with 1 method)
julia code_lowered(myf,(Array,))
1-element Array{Any,1}:
:($(Expr(:lambda, {:x},
{{:#s36,:#s40},{{:x,:Any,0},{:#s36,:Any,18},{:#s40,:Float64,2}},{}},
:(begin # none, line 1:
#s40 = top(typeassert)(top(convert)(Float64,getindex(x,1)),Float64)
#s36 = top(typeassert)(#s40,Float64) + 1
#s40 = top(typeassert)(top(convert)(Float64,#s36),Float64)
return #s36
end
julia myf_noassert(x) = x[1]+=1
myf_noassert (generic function with 1 method)
julia code_lowered(myf_noassert, (Array,))
1-element Array{Any,1}:
:($(Expr(:lambda, {:x}, {{:#s30},{{:x,:Any,0},{:#s30,:Any,18}},{}},
:(begin # none, line 1:
#s30 = getindex(x,1) + 1
setindex!(x,#s30,1)
return #s30
end
[julia-users] get is more eager than I thought
Is get supposed to evaluate the fallback argument? I was surprised by this:
julia dict1 = {1=1}
{1=1}
julia dict2 = Dict()
Dict{Any,Any}()
julia get(dict1,1,get!(dict2,1,2))
1
julia dict2
{1=2}
The other method is more what I'd have expected
julia get(dict1,1) do
()-get!(dict2,1,2)
end
1
julia dict2
Dict{Any,Any}()
