IceMorph morphological analysis data files
Data files
Jun 09, 2014 version files 17.39 MB
-
allvol.zip
839.94 KB
-
DataAvailabilityStatement.pdf
31.08 KB
-
dictionary_20140605.json
12.53 MB
-
icemorph_corpus_2014-06-01.zip
3.85 MB
-
IceMorph_data-readme.pdf
49.46 KB
-
tagged_corpora_20140605.json
81.15 KB
Abstract
This dataset consists of four main resources: a concatenated dictionary of Old Icelandic parsed for word class and inflectional detail; a corpus of Old Icelandic sagas in plain text and chunked by chapter; a tagged version of the same text, output of the IceMorph system; a training corpus labeled "Expert" for training and testing a machine learning module; and a training corpus labeled "Gold" for training and testing a machine learning module.
Methods
Datasets (1) dictionary (2a) saga texts were generated using OCR. Dataset (2b) is the output of the IceMorph tagging system. Datasets (3a) and (3b) were generated by hand-tagging.