textmodel_ca documentation

[vtmiller]

This is a really stupid issue, but I cannot figure out from the documentation what the "sv" variable in the ca-output is - can someone help please?

[kbenoit]

Can you show your output please?

[kbenoit]

You probably mean

> library("quanteda")
> dfmat <- dfm(tokens(data_corpus_irishbudget2010))
> tmod <- textmodel_ca(dfmat)
> summary(tmod)
           Length Class  Mode     
sv             7  -none- numeric  
nd             1  -none- numeric  
rownames      14  -none- character
rowmass       14  -none- numeric  
rowdist       14  -none- numeric  
rowinertia    14  -none- numeric  
rowcoord      98  -none- numeric  
rowsup         0  -none- logical  
colnames    5141  -none- character
colmass     5141  -none- numeric  
coldist     5141  -none- numeric  
colinertia  5141  -none- numeric  
colcoord   35987  -none- numeric  
colsup         0  -none- logical  
call           2  -none- call 

This is the default summary() method applied to the output, which is a an object defined by the ca package.

It looks a bit better when that package is attached:

> library("ca")
> library("quanteda")
> dfmat <- dfm(tokens(data_corpus_irishbudget2010))
> tmod <- textmodel_ca(dfmat)
> summary(tmod)

Principal inertias (eigenvalues):

 dim    value      %   cum%   scree plot               
 1      0.185186  19.5  19.5  *****                    
 2      0.139485  14.7  34.1  ****                     
 3      0.134083  14.1  48.2  ****                     
 4      0.127264  13.4  61.6  ***                      
 5      0.125959  13.2  74.9  ***                      
 6      0.121439  12.8  87.6  ***                      
 7      0.117473  12.4 100.0  ***                      
        -------- -----                                 
 Total: 0.950890 100.0                                 


Rows:
     name   mass  qlt  inr    k=1 cor ctr    k=2 cor ctr    k=3 cor ctr     k=4 cor ctr    k=5 cor ctr     k=6 cor ctr     k=7
1  | LBFF |  159 1000   78 |  676 852 392 | -162  49  30 |  -15   0   0 |   -24   1   1 | -119  26  18 |    74  10   7 |  -180
2  | BRFG |   82 1000   76 | -335  84  50 |  403 122  95 |  546 224 181 |   816 501 428 | -133  13  11 |    44   1   1 |  -268
3  | BJLA |  117 1000   80 | -413 162 108 | -905 777 689 |  190  34  32 |    73   5   5 |  118  13  13 |   -92   8   8 |   -15

...

Columns:
               name   mass  qlt  inr     k=1 cor ctr     k=2 cor ctr     k=3 cor ctr     k=4 cor ctr     k=5 cor ctr     k=6 cor
1    |         when |    2 1000    0 |  -379 501   1 |    74  19   0 |  -100  35   0 |  -281 275   1 |  -107  40   0 |  -149  77
2    |            i |    5 1000    1 |   106  44   0 |  -163 105   1 |   175 121   1 |  -359 508   5 |  -109  47   0 |  -190 142
3    |       prsntd |    0 1000    0 |   132  11   0 |    52   2   0 |  -801 401   0 |   -67   3   0 |  -866 468   0 |   271  46

See the ca package for things you can do with the output object. We don't have many post-model fitting functions in quanteda.textmodels for CA at the moment.