text
stringlengths 0
203
|
|---|
Tokens |
203,621 |
51,362 |
46,435 |
English data |
Training set |
Development set |
Test set |
LOC |
7140 |
1837 |
1668 |
MISC |
3438 |
922 |
702 |
ORG |
6321 |
1341 |
1661 |
PER |
6600 |
1842 |
1617 |
German data |
Training set |
Development set |
Test set |
Articles |
553 |
201 |
155 |
Sentences |
12,705 |
3,068 |
3,160 |
Tokens |
206,931 |
51,444 |
51,943 |
German data |
Training set |
Development set |
Test set |
LOC |
4363 |
1181 |
1035 |
MISC |
2288 |
1010 |
670 |
ORG |
2427 |
1241 |
773 |
PER |
2773 |
1401 |
1195 |
Table 1: Number of articles, sentences and tokens in |
each data file. |
Table 2: Number of named entities per data file |
2.3 |
between August 1996 and August 1997. For the |
training and development set, ten days’ worth of data |
were taken from the files representing the end of August 1996. For the test set, the texts were from December 1996. The preprocessed raw data covers the |
month of September 1996. |
The text for the German data was taken from the |
ECI Multilingual Text Corpus3 . This corpus consists |
of texts in many languages. The portion of data that |
was used for this task, was extracted from the German newspaper Frankfurter Rundshau. All three of |
the training, development and test sets were taken |
from articles written in one week at the end of August 1992. The raw data were taken from the months |
of September to December 1992. |
Table 1 contains an overview of the sizes of the |
data files. The unannotated data contain 17 million |
tokens (English) and 14 million tokens (German). |
2.2 |
Data preprocessing |
Subsets and Splits