TRAIN: twitter
TEST: cc
CROSS_TEST: True
eng cc cxg2
             precision    recall  f1-score   support

         au       0.33      0.66      0.44      5000
         ca       0.49      0.27      0.35      5000
         ch       0.12      0.80      0.20      2688
         ie       0.78      0.52      0.62      5000
         in       0.38      0.90      0.54      5000
         my       0.98      0.05      0.10      5000
         ng       0.73      0.41      0.52      5000
         nz       0.64      0.40      0.49      5000
         ph       0.87      0.22      0.35      5000
         pk       0.88      0.36      0.51      5000
         pt       0.22      0.17      0.19      3788
         uk       0.48      0.44      0.46      5000
         us       0.83      0.16      0.27      5000
         za       0.68      0.31      0.43      5000

avg / total       0.62      0.40      0.40     66476

[[3287   43 1443   15  137    0    1   36    0   15    6   16    0    1]
 [1001 1354 1630   64  410    0   20  255    1    9   27  174   23   32]
 [ 193   21 2143  131  134    0   18   12    0    8    5   14    4    5]
 [ 495   22  737 2604  359    0  206   57    1    4    6  453    1   55]
 [  89   33  148   15 4515    0   18   49    5   52   11   28   10   27]
 [ 517  158  852   50 1794  253  155  185   13   88   89  634   25  187]
 [ 283   44  515   33  257    0 2041   30    1   13 1501  118    8  156]
 [1158   74  817   33  482    0   10 1982    0    0   23  387    2   32]
 [ 230  249 1886   34  841    2   40   55 1105   20  414   61   46   17]
 [  55   29 1732    5 1232    1   14    8    4 1803   77   10    7   23]
 [  58   93 2696   24   28    1    3   23  132    9  654   22   42    3]
 [ 919   61 1037  103  491    0   53   77    0    4    4 2209    0   42]
 [1280  496  834  196  407    2  228  247    2   24   79  239  801  165]
 [ 436   73 1790   33  785    0    8   67    0    0   29  213    0 1566]]