selcukorkmaz / mvn Goto Github PK

Hi,

I want to conduct a mardia test on my data, which consists of different log transformed brain volumes. However, I get the following error. What does it mean and what can I do?

> Abide_Clean_All_log
# A tibble: 634 x 39
   Intra_Cranial_V… Total_Brain_Vol… CorticalWhiteMa… CSF_log[,1] Caudate_log[,1] Left_Thalamus_P…
              <dbl>            <dbl>            <dbl>       <dbl>           <dbl>            <dbl>
 1           0.602            0.604            -0.133      -0.826          -1.28           -0.173 
 2          -1.28            -1.28             -0.746      -0.570          -1.60           -1.46  
 3           1.32             1.32              0.929      -0.315           3.70            0.896 
 4           0.928            0.927             1.24        0.736          -1.15            1.70  
 5           0.0228           0.0234           -0.118      -0.198          -1.69           -0.820 
 6          -0.264           -0.265            -0.689       0.600           0.192          -1.03  
 7          -0.408           -0.406            -0.440      -1.39           -0.292           0.0737
 8           0.771            0.770             1.06        0.682          -2.83           -0.0949
 9           0.898            0.900             0.164      -1.20            1.46            0.846 
10           0.717            0.715             1.13        1.03            0.312           0.178 
# … with 624 more rows, and 33 more variables: Left_Putamen_log[,1] <dbl>,
#   Left_Pallidum_log[,1] <dbl>, Left_Hippocampus_log[,1] <dbl>,
#   Left_Cerebellum_Cortex_log[,1] <dbl>, Left_Caudate_log[,1] <dbl>, Left_Amygdala_log[,1] <dbl>,
#   Left_Accumbens_area_log[,1] <dbl>, Right_Thalamus_Proper_log[,1] <dbl>,
#   Right_Putamen_log[,1] <dbl>, Right_Pallidum_log[,1] <dbl>, Right_Hippocampus_log[,1] <dbl>,
#   Right_Cerebellum_Cortex_log[,1] <dbl>, Right_Caudate_log[,1] <dbl>,
#   Right_Amygdala_log[,1] <dbl>, Right_Accumbens_area_log[,1] <dbl>,
#   Left_CortexVol_log[,1] <dbl>, Right_CortexVol_log[,1] <dbl>, lhCortexVol_log[,1] <dbl>,
#   rhCortexVol_log[,1] <dbl>, CortexVol_log[,1] <dbl>, Left_VentralDC_log[,1] <dbl>,
#   Right_VentralDC_log[,1] <dbl>, Brain_Stem_log[,1] <dbl>, TotalGrayVol_log[,1] <dbl>,
#   Accumbens_area_log[,1] <dbl>, Amygdala_log[,1] <dbl>, Pallidum_Log[,1] <dbl>,
#   Putamen_log[,1] <dbl>, Hippocampus_log[,1] <dbl>, Thalamus_Proper_log[,1] <dbl>,
#   Cerebellum_Cortex_log[,1] <dbl>, VentralDC_log[,1] <dbl>, SUB_ID <dbl>

> mvn(Abide_Clean_All_log, mvnTest="mardia")
Error in solve.default(S, tol = tol) : 
  Lapack routine dgesv: system is exactly singular: U[38,38] = 0

Thank you for your help!
Camille

I just installed MVN and tried to run mvn() but got the following error:

MVN::mvn(preds)

Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]) : 
  there is no package called ‘forcats’

Once I manually installed forcats, the function worked fine. I can't see forcats as a dependency in the MVN DESCRIPTION file so I just thought I'd let you know.

This seems like a potentially useful package, but at this point I cannot use it for identifying multivariate outliers because the output using the arguments for multivariateOutlierMethod and showNewData does not include a way to link back the to the original data. Rather than the output of showNewData automatically filtering out the outliers, what I need is to be able to assess each case that is labeled as outlier by mvn(), which means having a way to go to the original observations for further examination. If you could add an index to the observations in the original data as part of the output that would be great. Thanks!

Hi!

I'm trying to use Royston test (via accumptions_manova function, but it uses MVN). And I get Error in royston(data, tol = tol) : n must be greater than 3.
I use columns "mean_normalized_focus_score", "mean_productivity_rating", "temptations_resistance", "temptations_challenge", "temptations_strength", "temptations_number". There are for sure more then 3 samples (you can check the data). The function works for other columns.
Do you know what might be an issue?
Thanks a lot!

Best,
Anastasia

data.csv

Hi,

After installing the MVN package, I am getting the below error on loading it. Can't figure the reason at the moment. This issue persists in both mac and windows.
Thanks

sROC 0.1-2 loaded
Error: package or namespace load failed for ‘MVN’:
.onLoad failed in loadNamespace() for 'rjags', details:
call: dyn.load(file, DLLpath = DLLpath, ...)
error: unable to load shared object '/Library/Frameworks/R.framework/Versions/3.5/Resources/library/rjags/libs/rjags.so':
dlopen(/Library/Frameworks/R.framework/Versions/3.5/Resources/library/rjags/libs/rjags.so, 10): Library not loaded: /usr/local/lib/libjags.4.dylib
Referenced from: /Library/Frameworks/R.framework/Versions/3.5/Resources/library/rjags/libs/rjags.so
Reason: image not found

I'm looking for advice to use MVN for large sets of data.

• Is there a method that supports memory allocation?
• or is there a way to use MVN in chunks?

results <- mvn(data = train, multivariateOutlierMethod = "quan", showOutliers = TRUE)

Error: cannot allocate vector of size 213.9 Gb