Distributional Part-of-Speech Tagging

By Hinrich Schitze. 1995

This paper presents an algorithm for tagging words whose part-of-speech properties are unknown. Unlike previous work, the algorithm categorizes word tokens in context instead of word types. The algorithm is evaluated on the Brown Corpus.

Subfields: Natural Language Processing, Computational Linguistics & Semantics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

In Computational Linguistics & Semantics

Citations in the corpus

Natural Language Processing (Almost) from Scratch (2011, methodology): “The induced word representation has been used with success in a wide variety of NLP tasks, including POS (Schütze, 1995), NER (Miller et al., 2004; Ratinov and Roth, 2009), or parsing (Koo et al., 2008).”
WordICA—emergence of linguistic representations for words by independent component analysis (2010, methodology): “Sch¨utze (1995) presented a method for tag induction in which the results of SVD were clustered into 200 distinct classes and further compared with approximately the same number of POS tags used in the Brown corpus (Francis and Kucera 1964).”
The effects of part–of–speech tagging on text–to–speech synthesis for resource–scarce languages (2010, background): “A popular approach to unsupervised learning of POS categories is the distributional tagging algorithm in [66], which can tag a language for which no knowledge about the categories is available beforehand.”
Word Embeddings through Hellinger PCA (2013, methodology): “Popular approaches such as Brown clustering algorithm (Brown et al., 1992) have been used with success in a wide variety of NLP tasks (Sch¨utze, 1995; Koo et al., 2008; Ratinov and Roth, 2009).”
Natural Language Grammar Induction Using a Constituent-Context Model (2001, background): “Several different distribu tional clustering approaches have resulted in relatively high-quality clusterings, though t he clusters’ resemblance to classical parts-of-speech varies substantially [9, 15].”
Learning syntactic structure (2007, methodology): “Using only the most frequent words in the corpus as contexts, various clustering methods have been used to induce part-of-speech tags (Schütze 1995; Clark 2000).”
Learning Syntactic Categories Using Paradigmatic Representations of Word Context (2012, methodology): “The ﬁrst subgroup represents each word type with its context vector and clusters these vectors accordingly (Sch¨utze, 1995).”
A fully Bayesian approach to unsupervised part-of-speech tagging (2007, methodology): “Distributional clustering and dimensionality reduction techniques are typically applied when linguistically meaningful classes are desired (Sch¨utze, 1995; Clark, 2000; Finch et al., 1995); probabilistic models have been used to ﬁnd classes that can improve smoothing and reduce perplexity (Brown et…”
Corpus-Based Induction of Syntactic Structure: Models of Dependency and Constituency (2004, methodology): “Previous work has shown that even this quite simple representation allows the induction of quite high quality word classes, largely corresponding to traditional parts of speech (Finch, 1993; Sch¨utze, 1995; Clark, 2000).”