Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. 1992. A practical part-of-speech tagger. In Proceedings of ANLP-92, pages 133--140, Trento, Italy.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....tend to be smaller. In work related to this [9] a number of taggers were applied to the problem of tagging Slovene. Four different taggers were trained and tested on a hand annotated corpus of Slovene, the translation of the novel 1984 by G. Orwell. The taggers tested were the HMM tagger [6, 15], Brill s Rule based tagger [3] the Maximum Entropy Tagger [14] and the Memory based Tagger [7] Accuracies on known words were mostly a little over 90 , with the Memory Based Tagger achieving 93.58 . Known words are those found in a lexicon that accompanies the corpus. Our goal here was to ....

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, pages 133-140, Trento, Italy, 1992.

.... part of speech taggers, the maximiza tion of Equation (1) to get the most likely tag se quence, is accomplished by the Viterbi algorithm (Church, 1988) and the maximum likelihood estimates of the parameters of Equation (2) are obtained frorn untagged corpus by the Forward Backward algorithm (Cutting et al. 1992). How ever, it is impossible to apply the Viterbi algorithm and the Forward Backward algorithm for word segmentation of those languages that have no delimiter between words, such as Japanese and Chinese, because word segmentation hypotheses overlap one another. Figure 3 shows an example of ....

Doug Cutting, Julian Ku- piec, Jan Pealersen, and Penelope Sibun. 1992. A Practical Part-of-Speech Tagger, In Proceedings of ANLP-92, pages 133-140.

.... to laboriously hand crafting rules for tagging, as was done in the past [Klein and Simmons, 1963; Harris, 1962] Almost all of the work in the area of automatically trained taggers has explored Markov model based part of speech tagging [Jelinek, 1985; Church, 1988; Derose, 1988; DeMarcken, 1990; Cutting et al. 1992; Kupiec, 1992; Chaxniak et al. 1993; Weischedel et al. 1993; Schutze and Singer, 1994; Lin et al. 1994; Elworthy, 1994; Merialdo 1995] 2 For a Markov model based tagger, training consists of learning both lexical probabilities (P(wordltag) and contextual probabilities (P(tagiltagi 1 . ....

.... bilities extracted from large manually annotated corpora (e.g. Weischedel et al. 1993; Charniak et al. 1993] It is possible to use unsupervised learning to train stochastic taggers without the need for a manually annotated corpus by using the Baum Welch al gorithm [Banm, 1972; Jelinek, 1985; Cutting et al. 1992; Kupiec, 1992; Elworthy, 1994; Merialdo, 1995] This algorithm works by iteratively adjusting the lexical and contextual probabilities to increase the overall probability of the training corpus. If no prior knowledge is available, probabilities are initially either assigned randomly or evenly ....

[Article contains additional citation context not shown here]

Cutting, D.; Kupiec, J.; Pedersen, J.; and Sibun, P. 1992. A prac- tical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, A CL, Trento, Italy.

....and an initial bias for the transition and output probabilities. The parameters are estimated by using the Baum Welch algorithm (Baum et al. 1970) The accuracy of the derived model depends heavily on the initial bias, but with a good choice results are comparable to those of method three (Cutting et M. 1992). This paper investigates a fifth method for estL mating natural language models, combining the advantages of the methods mentioned above. It is suitable for both speech recognition and partof speech tagging, has the advantage of automatically deriving word categories from a corpus and is ....

Doug Cutting, Julian Ku- piec, Jan Pedersen, and Penelope Sibun. 1992. A practical part-of-speech tagger. In Proceedings of the 3rd Uonference on Applied Natural Language Processing (ACL), pages 133-140.

....3 Sense Disambiguation with Adjacency Information The input to the disambiguator is a pair of words, along with the adjacency relationship that links them in the input text. The adjacency relationship is obtained automatically by processing the text through the Xerox PARC part of speech tagger [6] and a phrase extractor. The 12 adjacency relationships used by the disambiguator are listed below. These adjacency relationships were derived from an analysis of captions of news photographs provided by the Associated Press. The examples from the captions also helped us identify the heuristic ....

Cutting, Doug, Julian Kupiec, Jan Pedersen, and Penelope Sibun. 1992. "A Practical Part-of-Speech Tagger," in Proceedings of the Third Conference on Applied NLP.

....terms) corresponding to 41,842 concepts in SNOMED and 43,627 concepts in the Metathesaurus. Identifying adjectival modifiers The study of adjectival modification in the SNOMED terms under consideration was based on an underspecified syntactic analysis [5] that draws on a stochastic tagger [6] as well as the SPECIALIST Lexicon, a large syntactic lexicon of both general and medical English that is distributed with the UMLS. Although not perfect, this combination of resources effectively addresses the phenomenon of part of speech ambiguity in English, and, for example, correctly ....

Cutting DR, Kupiec J, Pedersen JO, Sibun P. A practical part-of-speech tagger. Proceedings of the Third Conference on Applied Natural Language Processing. 1992:133-140.

....HMM taggers have to be trained on previously tagged data in order to calculate the probabilities for tag sequences. Also, it has been demonstrated that with the use of the Expectation Maximisation (EM) algorithm (Dempster et al. 1 977) that stochastic models can be trained on untagged data (Cutting et al. 1992). However, this method still requires a dictionary of words with their respective word tags. The EM algorithm is then able to calculate the likelihood for each tag and tag transition probabilities. Experiments so far, however, have shown that taggers trained on tagged data will perform better than ....

D. Cutting, J. Kupiec, J. O. Pedersen and P. Sibun. A Practical part- of-speech tagger. In Third Conference on Applied Natural Language Processing, pp. 133-140. ACL. 1992.

....terms) corresponding to 41,842 concepts in SNOMED and 43,627 concepts in the Metathesaurus. 3.2. Identifying adjectival modifiers The study of adjectival modification in the SNOMED terms under consideration was based on an underspecified syntactic analysis [16] that draws on a stochastic tagger [17] as well as the SPECIALIST Lexicon, a large syntactic lexicon of both general and medical English that is distributed with UMLS. Although not perfect, this combination of resources effectively addresses the phenomenon of part of speech ambiguity in English, and, for example, correctly identifies ....

D.R. Cutting, J. Kupiec, J.O. Pedersen, P. Sibun, A practical part-of-speech tagger, in: Proceedings of the Third Conference on Applied Natural Language Processing, 1992, pp. 133#/140.

....colleagues [49, 50, 85, 84, 18] and Brill [15, 16] There have also been e#orts at learning parts of speech from word distributions, with application to tagging [76, 77] Taggers are currently wide spread and readily available. Those available for free include an HMM tagger implemented at Xerox [23], the Brill tagger, and the Multext tagger [8] Moreover, taggers have now been developed for a number of di#erent languages. Taggers have been described for Basque [6] Dutch [24] French [18] German [30, 75] Greek [24] Italian [24] Spanish [57] Swedish [13] and Turkish [63] to name a ....

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. A practical part-of-speech tagger. In Third Conference on Applied Natural Language Processing (ANLP-92), pages 133--140, 1992.

....can also be specified as well as feature negation. The optional ordering relations specify the internal ordering constraints among the dependents and regent. A sequence of tagged lexical items constitutes the input to the dependency parser. The tagger used was the Xerox part of speech tagger XPOST [3], modified for Swedish [11] Fig. 2 shows a typical dependency analysis of the sentence in Fig. 1. Identifying candidate phonological phrase boundaries using this representation as input is straight forward: they are found following each regent word (with the exception of the preposition infor ....

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. A practical part-of-speech tagger. In Proc. of the 3 rd Conference on Applied Natural Language Processing, pages 133--140, Trento, Italy, 1992.

....lexical interface, and the application of the parser to part of speech (POS) tagging for Norwegian. Applied to a small test set of about 4 000 words this method correctly tags 96 of the known words, which is close to the performance of other POS taggers trained on large text databases [2] [3] 85 of the unknown words are tagged correctly, and the probability of choosing the wrong pronunciation of a word from lexicon is less than 0.1 . 1. INTRODUCTION A parser in a text to speech (TTS) system should be fast and robust enough to make an analysis, no matter how ill formed the ....

....reasons for choosing this parsing algorithm for our TTS system. One possible application of the GLR parser is in POS tagging. Several methods for POS tagging or homograph disambiguation have been proposed the last years. Two of the most popular and well known are the Xerox stochastic POS tagger [2], and E. Brills rule based POS tagger [3] Both taggers apply manually POS tagged text corpora to induce a lexicon, and for testing. Today there exists no large tagged Norwegian corpus, and therefore we propose to use general lexical and grammatical information available in a lexicon and a grammar ....

[Article contains additional citation context not shown here]

Cutting, D. et. al., "A practical part-of-speech tagger.", Proc. 3. Conf. .on Applied NLP, Trento, Italy, 1992.

....should be similar to those of English i.e. Another problem for these languages is also related to segmentation: Part of Speech (POS) tagging. Taking Chinese as an example, although there are many Part of Speech taggers available for English and these taggers are claimed to be language independent [9, 12, 13, 17, 27, 51, 76], there is no publicly available POS tagger for Chinese. The main reason for this is the lack of benchmark data for training and testing. With the recent release of Chinese Treebank [22] we are now able to port the taggers to Chinese. Then with a Chinese segmenter and a POS tagger, we can build ....

....information is useful in many NLP task including information retrieval, speech recognition, language modeling, text to speech synthesis. In the MaxEnt name entity recognition system, we are also going to use the POS information. Although there are some publicly available POS tagger for English, [9, 12, 13, 17, 27, 51, 76], and they are claimed to be language independent, there is no such a POS tagger for Chinese. It is easy to port these POS tagger from English to other western language, such as Spain and French, because these languages are all similar in the sense that: 35 1. Each character is represented with ....

[Article contains additional citation context not shown here]

Cutting, D. et al. A Practical Part-of-Speech Tagger In Proceedings of the Third Conference on Applied Natural Language Processing, 1992.

....disambiguation relies on the fact that certain sequences of parts of speech are more probable than others. Generally, the frequency of sequences of tags in hand tagged texts are used to estimate these probabilities. 1 Ftp able at clarity.princeton.edu 2 In Hidden Markov Model (HMM) tagging [3], the probability of seeing a given tag depends on the tags preceding it. The HMM training and tagging programs in our experiment [9] is based bigrams, i.e. it takes into account only the immediate context of a word. The probabilities in the HMM model are based on ambiguity classes. An ambiguity ....

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun, `A Practical Part-of-speech Tagger', in Proceedings of ANLP-92, Trento, Italy, (1992).

....H. Paulussen and W. Martin 1992; A. Voutilainen 1994; Ph. Laval 1995) This scheme is opposed to another, in which the distributional data is automatically acquired by frequency based corpus training (I. Marshall 1983; F. Jelinek 1985; R. Garside 1987; J. Benello et al. 1989; D. Hindle 1989; D. Cutting et al. 1992; E. Brill 1992; H. Schmid 1994; E. Roche and Y. Schabs 1995) Building a grammar of resolution of lexical ambiguity is a challenge: correct analyses 3 should not be removed; the results of syntactic parsing cannot be explicitly used, since it is not available at the time when ambiguity ....

Cutting, D.; J. Kupiec; J. Pedersen; P. Sibun. 1992. A practical part-of-speech tagger. In Q...'prrqvt+'s#ur"q8'sr...rpr'6ffyvrqIh#...hyGhthtrQ...'pr++vt, Trento, ACL, pp.133-140.

....tagger and the local grammar systems) A rule based approach has also been advocated within the Dutch TOSCA project. 4.2.2 Probabilistic Tagging Methods 4.2.2. 1 Short overview The majority of part of speech taggers ( Bahl and Mercer, 1976] Church, 1988] Leech et al. 1994] DeRose, 1988] [Cutting et al. 1992], Meteer et al. 1991] Wothke et al. 1993] Feldweg, 1993] Kempe, 1993] Kempe, 1994] and [Schmid, 1994a] make use of probabilistic methods. Language is modeled as a Markov process of k th order, where k is for most taggers either 1 (bigram tagger) or 2 (trigram tagger) The probability ....

....developed for large morphosyntactic tagsets, where estimates are particularly hard to do. In this latter approach, tags are split into components and the contextual probabilities are deduced from component probabilities. It is also possible to estimate model parameters from an untagged corpus [Cutting et al. 1992]. An initial set of model parameters is used to tag the training corpus. Then an improved set of parameters is estimated from the corpus. The corpus is tagged again and so on, until the accuracy does not improve anymore. The forward backward algorithm [Baum, 1972] is used for this reestimation ....

Cutting, D., Kupiec, J., Pedersen, J., and Sibun, P. (1992). A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, pages 133--140.

.... to automatic PoS tagging: rule based approaches [1] use linguistic knowledge to formulate simple rules that assign a part of speech to an ambiguous word using context information; statistical approaches (of which hidden Markov models trained using the Baum Welch expectationmaximization algorithm [2, 13] [10, ch. 10] are the standard model) use the statistics collected from ambiguously or unambiguously tagged texts (see below) to estimate the likelihood of each possible interpretation of a sentence or text portion so that the most likely disambiguation is chosen. Of course, hybrid approaches are ....

.... PoS tag (unknown words are usually assigned the set of open categories, that is, categories to which it is very possible to add new words of the language: nouns, verbs, adjectives, adverbs and proper nouns) Words receiving the same set of PoS tags are said to belong to the same ambiguity class [2]; for example, the words tailor and book belong to the ambiguity class fnoun, verbg. A neural or connectionist approach is also possible; a brief survey of neural PoS tagging work follows: Schmid [14] trains a single layer perceptron to produce the PoS tag of a word as a unary or onehot ....

[Article contains additional citation context not shown here]

Doug Cutting, Julian Kupiec, Jan Pedersen and Penelope Sibun, \A Practical Part-of-Speech Tagger", Proceedings of Third Conference on Applied Natural Language Processing, Association for Computational Linguistics, 1992, pp. 133-140.

....Co Training, which has been used successfully in several classification tasks like web page classification, word sense disambiguation and named entity recognition. Early work in combining labeled and unlabeled data for NLP tasks was done in the area of unsupervised part of speech (POS) tagging. (Cutting et al. 1992) reported very high results (96 on the Brown corpus) for unsupervised POS tagging using Hidden Markov Models (HMMs) by exploiting hand built tag dictionaries and equivalence classes. Tag dictionaries are predefined assignments of all possible POS tags to words in the test data. This impressive ....

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. 1992. A practical part-of-speech tagger. In Proc. of 3rd ANLP Conf., Trento, Italy. ACL.

.... the Viterbi algorithm (Viterbi, 1967; Jelinek et al., 1975) Recently, the Baum Welch algorithm (Baum, 1972) a self organising technique for hidden Markov models which can be applied to ambiguous training data (Jelinek, 1985) has been utilised for part of speech tagging (Kupiec, 1992; Cutting et al., 1992). In this latter model, lexical probabilities for all but the most frequent words are replaced by equivalence 1 classes of words assigned the same ambiguous set of tags, and bigram transition probabilities are augmented with manually specified networks of tied states to overcome observed common ....

....the construction of such corpora is labour intensive. Meteer et al. (1991) estimate that around 70K words of such training material is required to produce an accurate trigram tagger if direct estimates of parameters are required (given the extreme bias in distribution of observed trigrams) whilst Cutting et al. (1992) mention that they have achieved reasonable results using iterative re estimation on 3000 untagged sentences. However, both these lower limit estimates of training corpus size assume that some other accurate method is employed to estimate lexical probabilities or ambiguity class membership for ....

[Article contains additional citation context not shown here]

Cutting, D., Kupiec, J., Pedersen, J., and Sibun, P. (1992). "A practical part-of-speech tagger." In Proceedings, 3rd Applied ACL. 133--140.

.... It contains approximately 27,000 articles, which are accessed via the Text Database (TDB) Cutting et al. 1991] which is a flexible platform for the development of retrieval system prototypes and is struc tured so that additional functional components (e.g. search strategies and text taggers [Cutting et al. 1992]) can be easily integrated. The components responsible for linguistic analysis are a part of speech tagger and a lexico syntactic pattern marcher. The tagger is based on a hidden Markov model (HMM) HMM s are probabilistic and their parameters can be estimated by training on a sample of ordinary ....

D. Cutting, J. Kupiec, J. Peder- sen, and P. Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, Trento, Italy, April 1992. ACL.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. 1992. A practical part-of-speech tagger. In Proceedings of ANLP-92, pages 133--140, Trento, Italy.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. A practical part-of-speech tagger. In Proc. of the Third Conf. on Applied Natural Language Processing. Association for Computational Linguistics, 1992.

No context found.

D. Cutting, J. Kupiec, J. Pederson and P. Sibun, "A practical part-of-speech tagger". Proceeding on Third Conference on Applied Natural Language Processing. Trento, Italy, 1992.

No context found.

Cutting, Doug, Julian Kupiec, Jan Pederson, and Penelope Sibun. 1992. A practical part-of-speech tagger. In Third Conference on Applied Natural Language Processing, pages 133#140, Trento, Italy.

No context found.

Douglas R. Cutting, Julian M. Kupiec, Jan O. Pedersen, and Penelope Sibun. A Practical Part-of-Speech Tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, pages 133--140, Trento, Italy, April 1992.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, 1992. 94

No context found.

D. Cutting et al. 1992. A practical part-of-speech tagger. In Proc. 3rd Conf. on Applied Natural Language Processing.

No context found.

D. Cutting et al. 1992. A practical part-of-speech tagger. In Proc. 3rd Conf. on Applied Natural Language Processing.

No context found.

Cutting D., J Kupiec., Pederson J., Sibun P. (1992) A practical part-of-speech tagger. Proceeding on Third Conference on Applied Natural Language Processing. Trento, Italy.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, P. Sibun. A practical part-of-speech tagger. Proc. 3

No context found.

Cutting DR, Kupiec J, Pedersen JO, Sibun P. A practical part-of-speech tagger. Proceedings of the Third Conference on Applied Natural Language Processing, 1992.

No context found.

Cutting D, Kupiec J, Pedersen J and Sibun P. A practical part-of-speech tagger. Proc Third Conference on Applied Natural Language Processing, 1992.

No context found.

Cutting D, Kupiec J, Pedersen J, et al. A practical part-of-speech tagger. Proc Conf Applied NLP. 1992;:133-40.

No context found.

Cutting D.; J. Kupiec; J. Pedersen; and P. Sibun. 1992. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing.

No context found.

Cutting D, Kupiec J, Pedersen J and Sibun P. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, 1992.

No context found.

Cutting D, Kupiec J, Pedersen J and Sibun P. A practical part-of-speech tagger. Proceedings of the Third Conference on Applied Natural Language Processing, 1992

No context found.

Cutting, D., Kupiec, J., Pedersen, J., Sibun, P.: A practical part-of-speech tagger. In: Proceedings of the Third Conference on Applied Natural Language Processing, Trento, Italy (1992) 133--140.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun, "A Practical Part-of-Speech Tagger, " in Proc. the Third Conference on Applied Natural Language Processing, pp. 133-140, 1992.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. 1992. A Practical Part-of-Speech Tagger. In Proc. of the ANLP'92.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing ANLP-92, 1992.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, 1992.

No context found.

D. Cutting, J. Kupiec, J. Pedersen, and P. Sibun. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, Trento, Italy, 1992.

No context found.

Cutting D., Kupiec J., Pedersen J., Sibun P. A Practical Part-of-Speech Tagger. In Proceedings of the 3rd conference on Applied Natural Language Processing (ANLP). 1992. ftp://parcftp.xerox.com/pub/tagger/

No context found.

Cutting, D., Kupiec, J., Pedersen, J. and Sibun, P. (1992). A Practical Part-of-speech Tagger. In Third Conference on Applied Natural Language Processing, ACL, 133--140.

No context found.

Cutting, D., Kupiec, J., Pederson, J. and Sibun, P. A Practical Part-of-Speech Tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, ACL, 1992.

No context found.

Cutting, D., Kupiec, J., Pedersen, J., Sibun, P. (1992). A Practical Part of Speech Tagger, Proceedings of the Third International Conference on Applied Natural Language Processing, Trento, 133-140.

No context found.

Cutting, D., Kupiec, J., Pederson, J. and Sibun, P.A Practical Part-of-SpeechTagger. In Proceedings of the Third Conference on Applied Natural Language Processing, ACL, 1992.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen and Penelope Sibun. A Practical Part-of-Speech Tagger. In Third Conference on Applied Natural Language Processing. pages 133--140. Trento, 1992.

No context found.

Cutting, D., J. Kupiec, J. Pedersen, and P. Sibun, 1992. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing. Trento, Italy.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen, and Penelope Sibun. 1992. A practical part-of-speech tagger. In Proceedings of the Third Conference on Applied Natural Language Processing, Trento, Italy.

No context found.

Doug Cutting, Julian Kupiec, Jan Pedersen & Penelope Sibun, "A Practical Part-of-Speech Tagger," Proceedings of the Third Conference on Applied Natural Language Processing (1992). 127

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC