Improving Data Driven Part-of-Speech Tagging by Morphologic Knowledge Induction
Uwe D. Reichel · Open access LMU (Ludwid Maxmilian's Universitat Munchen) · 2005
Abstract. We present a Markov part-of-speech tagger for which the P(w|t) emission probabilities of word w given tag t are replaced by a linear interpolation of tag emission probabilities given a list of representations of w. As word representations, string suffixes of w are cut off at the local maxima of the Normalized Backward Successor Variety. This procedure allows for the derivation of linguistically meaningful string suffixes that may relate to certain POS labels. Since no linguistic knowledge is needed, the procedure is language independent. Basic Markov model part-of-speech taggers are significantly outperformed by our model. 1