A Multilingual Evaluation of Three Spelling Normalisation Methods for Historical Text

Pettersson, Eva; Megyesi, Beáta; Nivre, Joakim

doi:10.3115/v1/w14-0605

Cited by 27 publications

(43 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Table 9 shows results reported by Pettersson et al (2014) on English, German, Hungarian, Icelandic and Swedish, and results by Sánchez-Martínez et al (2013) on Spanish, in addition to our results on Slovene. However, the experimental setups are fairly difficult to compare.…”

Section: Cross-language Comparisonssupporting

confidence: 68%

“…It corresponds to our Baseline 2, to some similar experiments but with language-specific thresholds for Pettersson et al (2014), and to a comparable approach based on a spellchecker for Sánchez-Martínez et al (2013). Again, this method shows comparatively low accuracy values for Slovene, but similar CER values.…”

Section: Cross-language Comparisonsmentioning

confidence: 99%

“…column shows the results of CSMT using gold standard training data, i.e., our supervised setting. We show the best values of all giza or m2m experiments of Pettersson et al (2014), the SMT experiment of Sánchez-Martínez et al (2013), and the best setting in terms of accuracy as given in Fig. 2.…”

Section: Cross-language Comparisonsmentioning

confidence: 99%

See 2 more Smart Citations

Modernising historical Slovene words

Scherrer¹,

Erjavec²

2015

Nat. Lang. Eng.

View full text Add to dashboard Cite

We propose a language-independent word normalisation method and exemplify it on modernising historical Slovene words. Our method relies on character-level statistical machine translation (CSMT) and uses only shallow knowledge. We present relevant data on historical Slovene, consisting of two (partially) manually annotated corpora and the lexicons derived from these corpora, containing historical word-modern word pairs. The two lexicons are disjoint, with one serving as the training set containing 40,000 entries, and the other as a test set with 20,000 entries. The data spans the years 1750-1900, and the lexicons are split into 50-year slices, with all the experiments carried out separately on the three time periods. We perform two sets of experiments. In the first one -a supervised setting -we build a CSMT system using the lexicon of word pairs as training data. In the second one -an unsupervised setting -we simulate a scenario in which word pairs are not available. We propose a two-step method where we first extract a noisy list of word pairs by matching historical words with cognate modern words, and then train a CSMT system on these pairs. In both sets of experiments we also optionally make use of a lexicon of modern words to filter the modernisation hypotheses. While we show that both methods produce significantly better results than the baselines, their accuracy and which method works best strongly correlates with the age of the texts, meaning that the choice of the best method will depend on the properties of the historical language which is to be modernised. As an extrinsic evaluation we also compare the quality of part-of-speech tagging and lemmatisation directly on historical text and on its modernised words. We show that, depending on the age of the text, annotation on modernised words also produces significantly better results than annotation on the original text.

show abstract

Section: Cross-language Comparisonssupporting

confidence: 68%

Section: Cross-language Comparisonsmentioning

confidence: 99%

See 1 more Smart Citation

Modernising historical Slovene words

Scherrer¹,

Erjavec²

2015

Nat. Lang. Eng.

View full text Add to dashboard Cite

show abstract

“…Recently, work on non-standard historical varieties has focused on spelling normalization using rule-based, statistical and neural stringtransduction models (Pettersson et al, 2014;Bollmann and Søgaard, 2016;Tang et al, 2018). Previous studies on lemmatization of historical variants focused on evaluating off-the-shelf systems.…”

Section: Related Workmentioning

confidence: 99%

Improving Lemmatization of Non-Standard Languages with Joint Learning

Manjavacas¹,

Kádár²,

Kestemont³

2019

Proceedings of the 2019 Conference of the North

View full text Add to dashboard Cite

Lemmatization of standard languages is concerned with (i) abstracting over morphological differences and (ii) resolving token-lemma ambiguities of inflected words in order to map them to a dictionary headword. In the present paper we aim to improve lemmatization performance on a set of non-standard historical languages in which the difficulty is increased by an additional aspect (iii): spelling variation due to lacking orthographic standards. We approach lemmatization as a stringtransduction task with an encoder-decoder architecture which we enrich with sentence context information using a hierarchical sentence encoder. We show significant improvements over the state-of-the-art when training the sentence encoder jointly for lemmatization and language modeling. Crucially, our architecture does not require POS or morphological annotations, which are not always available for historical corpora. Additionally, we also test the proposed model on a set of typologically diverse standard languages showing results on par or better than a model without enhanced sentence representations and previous state-of-the-art systems. Finally, to encourage future work on processing of non-standard varieties, we release the dataset of non-standard languages underlying the present study, based on openly accessible sources.

show abstract

“…Data-driven approaches, on the other hand, derive their knowledge from large amounts of normalized data. A recent popular data-driven methodology for normalization is the use of character-based statistical machine translation (Schulz et al 2016, Pettersson et al 2014, which partly overcomes the requirement for large amounts of data as the system works at the character level.…”

Section: The Effects Of Normalization On Tagging Accuracymentioning

confidence: 99%

An automatic part-of-speech tagger for Middle Low German

Koleva

Farasyn

Desmet

et al. 2017

IJCL

View full text Add to dashboard Cite

Syntactically annotated corpora are highly important for enabling large-scale diachronic and diatopic language research. Such corpora have recently been developed for a variety of historical languages, or are still under development. One of those under development is the fully tagged and parsed Corpus of Historical Low German (CHLG), which is aimed at facilitating research into the highly under-researched diachronic syntax of Low German. The present paper reports on a crucial step in creating the corpus, viz. the creation of a part-of-speech tagger for Middle Low German (MLG). Having been transmitted in several non-standardised written varieties, MLG poses a challenge to standard POS taggers, which usually rely on normalized spelling. We outline the major issues faced in the creation of the tagger and present our solutions to them.

show abstract

A Multilingual Evaluation of Three Spelling Normalisation Methods for Historical Text

Cited by 27 publications

References 5 publications

Modernising historical Slovene words

Modernising historical Slovene words

Improving Lemmatization of Non-Standard Languages with Joint Learning

An automatic part-of-speech tagger for Middle Low German

Contact Info

Product

Resources

About