Evaluating an ‘off-the-shelf ’ POS-tagger on Early Modern German text
Silke Scheible, Richard J. Whitt, Martin Durrell, Paul Bennett · 2012
The goal of this study is to evaluate an ‘offthe-shelf’ POS-tagger for modern German on historical data from the Early Modern period (1650-1800). With no specialised tagger available for this particular stage of the language, our findings will be of particular interest to smaller, humanities-based projects wishing to add POS annotations to their historical data but which lack the means or resources to train a POS tagger themselves. Our study assesses the effects of spelling variation on the performance of the tagger, and investigates to what extent tagger performance can be improved by using ‘normalised ’ input, where spelling variants in the corpus are standardised to a modern form. Our findings show that adding such a normalisation layer improves tagger performance considerably.