[julia-users] Re: Chi square test in Julia?
Thanks Benjamin! Creating a matrix out of my two arrays works just fine. I'll go that way for now. Josh On Tuesday, November 10, 2015 at 4:32:27 PM UTC-6, Benjamin Deonovic wrote: > > Looks like the state of Julia's "Factors" (PooledDataArrays) is still > quite a fast evolving monster. Because of this there hasn't been a proper > implementation of R's "table" function. For now, if you want to run > ChisqTest in Julia on two vectors do it on the matrix you would obtain in R > by running table(x,y). >
[julia-users] Re: Chi square test in Julia?
I found your Chi-Square test and am trying to use it. It appears to work
with one array but not with two. My steps are below:
This works:
ChisqTest([1,2,3,4])
This doesn't:
ChisqTest([1,2,3,4],[1,2,2,4])
It errors with the following:
LoadError: MethodError: `ChisqTest` has no method matching
ChisqTest(::Array{Int64,1}, ::Array{Int64,1})
Closest candidates are:
ChisqTest{T<:Integer}(::AbstractArray{T<:Integer,1},
::AbstractArray{T<:Integer,1},
!Matched::Tuple{UnitRange{T<:Integer},UnitRange{T<:Integer}})
ChisqTest{T<:Integer}(::AbstractArray{T<:Integer,1},
::AbstractArray{T<:Integer,1}, !Matched::T<:Integer)
ChisqTest{T<:Integer,U<:AbstractFloat}(::AbstractArray{T<:Integer,1},
!Matched::Array{U<:AbstractFloat,1})
...
while loading In[118], in expression starting on line 1
I have checked the arrays to make sure they're AbstractArrays and the
result is true.
Any advice would be helpful, I might just be implementing wrong.
On Friday, March 6, 2015 at 7:58:20 AM UTC-6, Benjamin Deonovic wrote:
>
> My pull has been merged: https://github.com/JuliaStats/HypothesisTests.jl
>
> On Thursday, March 5, 2015 at 8:32:32 AM UTC-6, Benjamin Deonovic wrote:
>>
>> I implemented the chisquare test in julia. I made a pull request in the
>> HypothesisTests package. It hasn't been pulled yet, but probably will be
>> soon.
>>
>> On Sunday, February 8, 2015 at 5:32:48 PM UTC-6, Arin Basu wrote:
>>>
>>> Hi All,
>>>
>>> Please pardon my ignorance, but how does one do chisquare test in Julia.
>>> Something like,
>>>
>>> ```
>>>
>>> chisq.test(x, y = NULL, correct = TRUE,
>>>p = rep(1/length(x), length(x)), rescale.p = FALSE,
>>>simulate.p.value = FALSE, B = 2000)
>>>
>>> ```
>>>
>>> in R
>>>
>>>
>>> I could not find anything in the documentation. I must not have searched
>>> enough, what can it be?
>>>
>>>
>>> Best,
>>>
>>> Arin
>>>
>>>
[julia-users] Re: Chi square test in Julia?
Try a matrix?
[julia-users] Re: Chi square test in Julia?
Hey Joshua,
Just saw your post. I will investigate into what the issue is. I wrote this
quite a while ago when I was just learning Julia!
On Tuesday, November 10, 2015 at 9:10:10 AM UTC-6, Joshua Duncan wrote:
>
> I found your Chi-Square test and am trying to use it. It appears to work
> with one array but not with two. My steps are below:
>
> This works:
> ChisqTest([1,2,3,4])
>
> This doesn't:
> ChisqTest([1,2,3,4],[1,2,2,4])
>
> It errors with the following:
>
> LoadError: MethodError: `ChisqTest` has no method matching
> ChisqTest(::Array{Int64,1}, ::Array{Int64,1})
> Closest candidates are:
> ChisqTest{T<:Integer}(::AbstractArray{T<:Integer,1},
> ::AbstractArray{T<:Integer,1},
> !Matched::Tuple{UnitRange{T<:Integer},UnitRange{T<:Integer}})
> ChisqTest{T<:Integer}(::AbstractArray{T<:Integer,1},
> ::AbstractArray{T<:Integer,1}, !Matched::T<:Integer)
> ChisqTest{T<:Integer,U<:AbstractFloat}(::AbstractArray{T<:Integer,1},
> !Matched::Array{U<:AbstractFloat,1})
> ...
> while loading In[118], in expression starting on line 1
>
>
>
> I have checked the arrays to make sure they're AbstractArrays and the
> result is true.
>
> Any advice would be helpful, I might just be implementing wrong.
>
> On Friday, March 6, 2015 at 7:58:20 AM UTC-6, Benjamin Deonovic wrote:
>>
>> My pull has been merged: https://github.com/JuliaStats/HypothesisTests.jl
>>
>> On Thursday, March 5, 2015 at 8:32:32 AM UTC-6, Benjamin Deonovic wrote:
>>>
>>> I implemented the chisquare test in julia. I made a pull request in the
>>> HypothesisTests package. It hasn't been pulled yet, but probably will be
>>> soon.
>>>
>>> On Sunday, February 8, 2015 at 5:32:48 PM UTC-6, Arin Basu wrote:
Hi All,
Please pardon my ignorance, but how does one do chisquare test in
Julia. Something like,
```
chisq.test(x, y = NULL, correct = TRUE,
p = rep(1/length(x), length(x)), rescale.p = FALSE,
simulate.p.value = FALSE, B = 2000)
```
in R
I could not find anything in the documentation. I must not have searched
enough, what can it be?
Best,
Arin
[julia-users] Re: Chi square test in Julia?
Looks like the state of Julia's "Factors" (PooledDataArrays) is still quite a fast evolving monster. Because of this there hasn't been a proper implementation of R's "table" function. For now, if you want to run ChisqTest in Julia on two vectors do it on the matrix you would obtain in R by running table(x,y).
[julia-users] Re: Chi square test in Julia?
Okay I have figured out the issue. I will fix it so it works the way you expected it to work. Before the fix goes live though it should work to do: ChisqTest([1,2,3,4],[1,2,2,4], 4) *note the 4 The issue was when I submitted the code to HypothesisTests.jl the only way to create a contingency table between two vectors x and y was to also provide the levels that the categorical variables could take on. Afterwards I submited a version to StatsBase.jl for ``counts`` that had default values, but I forgot to update my code at HypothesisTests.jl. Sorry for the late response! **Note the above will give you what is equivalent in R to: > chisq.test(matrix(c(1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,1),nrow = 4,ncol = 4)) Pearson's Chi-squared test data: matrix(c(1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1), nrow = 4, ncol = 4) X-squared = NaN, df = 9, p-value = NA Warning message: In chisq.test(matrix(c(1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, : Chi-squared approximation may be incorrect However you are probably interested in > chisq.test(c(1,2,3,4),c(1,2,2,4)) Pearson's Chi-squared test data: c(1, 2, 3, 4) and c(1, 2, 2, 4) X-squared = 8, df = 6, p-value = 0.2381 When I wrote up the function there wasn't a good equivalent of R's ``table`` function. I will try to flesh out this code so it works closer to R. In the meantime the best thing is to work with the full contingency table i.e.: julia> ChisqTest([1 0 0; 0 1 0; 0 1 0; 0 0 1]) Pearson's Chi-square Test - Population details: parameter of interest: Multinomial Probabilities value under h_0: [0.0625,0.0625,0.0625,0.0625,0.125,0.125,0.125,0.125,0.0625,0.0625,0.0625,0.0625] point estimate: [0.25,0.0,0.0,0.0,0.0,0.25,0.25,0.0,0.0,0.0,0.0,0.25] 95% confidence interval: [(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0),(0.0,1.0)] Test summary: outcome with 95% confidence: fail to reject h_0 two-sided p-value: 0.23810330555354436 (not significant) Details: Sample size: 4 statistic: 8.0 degrees of freedom: 6 residuals: [1.5,-0.5,-0.5,-0.5,-0.7071067811865475,0.7071067811865475,0.7071067811865475,-0.7071067811865475,-0.5,-0.5,-0.5,1.5] std. residuals: [2.0,-0.,-0.,-0.,-1.1547005383792517,1.1547005383792517,1.1547005383792517,-1.1547005383792517,-0.,-0.66 66,-0.,2.0]
[julia-users] Re: Chi square test in Julia?
My pull has been merged: https://github.com/JuliaStats/HypothesisTests.jl On Thursday, March 5, 2015 at 8:32:32 AM UTC-6, Benjamin Deonovic wrote: I implemented the chisquare test in julia. I made a pull request in the HypothesisTests package. It hasn't been pulled yet, but probably will be soon. On Sunday, February 8, 2015 at 5:32:48 PM UTC-6, Arin Basu wrote: Hi All, Please pardon my ignorance, but how does one do chisquare test in Julia. Something like, ``` chisq.test(x, y = NULL, correct = TRUE, p = rep(1/length(x), length(x)), rescale.p = FALSE, simulate.p.value = FALSE, B = 2000) ``` in R I could not find anything in the documentation. I must not have searched enough, what can it be? Best, Arin
[julia-users] Re: Chi square test in Julia?
I implemented the chisquare test in julia. I made a pull request in the HypothesisTests package. It hasn't been pulled yet, but probably will be soon. On Sunday, February 8, 2015 at 5:32:48 PM UTC-6, Arin Basu wrote: Hi All, Please pardon my ignorance, but how does one do chisquare test in Julia. Something like, ``` chisq.test(x, y = NULL, correct = TRUE, p = rep(1/length(x), length(x)), rescale.p = FALSE, simulate.p.value = FALSE, B = 2000) ``` in R I could not find anything in the documentation. I must not have searched enough, what can it be? Best, Arin
[julia-users] Re: Chi square test in Julia?
Hi Arin, It would appear that there isn't a chi square test at the moment. If you want to implement one yourself, you can compute the quantiles of a chi squared distribution using DIstributions.jl (if you do this, please consider submitting it to HypothesisTests.jl). Alternatively, you could call the R function via RCall.jl or Rif.jl. SImon On Sunday, 8 February 2015 23:32:48 UTC, Arin Basu wrote: Hi All, Please pardon my ignorance, but how does one do chisquare test in Julia. Something like, ``` chisq.test(x, y = NULL, correct = TRUE, p = rep(1/length(x), length(x)), rescale.p = FALSE, simulate.p.value = FALSE, B = 2000) ``` in R I could not find anything in the documentation. I must not have searched enough, what can it be? Best, Arin
