Search

—This paper is the continuation of a work submitted to the International Conference Corpus Linguistics 2021 [1]. On that occasion, a rule-based stochastic hybrid part-of-speech tagger (POS) was introduced for Sranan Tongo, a Creole language from South America with around half a million speakers. Since Sranan Tongo does not have a written corpus and text annotation is an expensive and time-consuming task, it was proposed to take a first step in training a POS tagger using only 550 hand-annotated sentences with part of speech tags. In this new contribution, the development of the POS tagger for Sranan Tongo goes a step further with the addition of more training data. For this matter, the tagger was used to annotate 2,406 sentences. The tagging results were hand-corrected and employed to retrain the model. A comparison is shown between the performance of the POS tagger on three texts before and after the inclusion of the new training data., International Journal of Open Information Technologies, Выпуск 12 2022

Kada govorimo o računalnoj analizi i razumijevanju teksta, alati poput lematizatora, korjenovatelja, obilježivača vrsta riječi te različiti korpusi jezika igraju veliku ulogu u području računalne lingvistike. Alati poput ovih promatraju sintaksu i lingvistiku nekog određenog jezika te što boljom primjenom pravila istih, uz pokoju implementaciju vjerojatnosnih algoritama, nastoje bolje obraditi zadani jezik. U ovom završnom radu obrađujemo 4 najpoznatija korjenovatelja, dva lematizatora te dva obilježivača vrsta riječi njemačkog jezika. Osim teorijske obrade navedenih alata, dotaknut ćemo se i praktične usporedbe navedenih u zasebnom poglavlju na vlastitim tekstovima. Korjenovatelji Snowball, CISTEM, Text::Geramn i UniNE, lematizatori GermaLemma i SMOR te obilježivači TIGER korpusa i Pro3GreDE imaju iskazanu točnost u postotcima. Među korjenovateljima najuspješniji se pokazao CISTEM korjenovatelj s 91.23% točnih korjenovanja, zatim Text::German sa 88, 55% kojeg slijedi Snowball sa 82, 44% te na kraju UniNE koji ima točnost u rasponu od 78, 63% do 80, 92%. Između dva obilježivača vrsta riječi točniji se pokazao hibridni Pro3GresDE sa 93, 55% te onaj uključen unutar TIGER korpusa sa 90, 32% točnosti. Kod lematizatora točnijim se pokazao SMOR sa 94, 27% točnosti te nakon njega GermaLemma sa 85, 5% točnosti.

The purpose of this study was to analyze parts of speech used from selected news online entitled “Thailand's anti-smoking YouTube hit”. This study analyzed only words existed in selected news online from Bangkok Post, a Thai newspaper online. It focused on parts of speech used; noun, verb, adjective, adverb, pronoun, preposition, determination, conjunction and number. The researchers use only a website “parts-of-speech.info” to analyze and collect data. Selected news from Bangkok Post online entitled “Thailand's anti-smoking YouTube hit” was main material for this independent study. Parts-of-speech.info are a website used to collect and analyze data from the selected news. Parts-of-speech.info were used to analyze each sentence from the selected news. The results showed total of; adjective, adverb, conjunction, determiner, noun, number, preposition, pronoun and verb that existed in each sentence. The findings may be useful for English teachers who would like to teach all skills using online news as their teaching resources.

International audience; This work investigates the feasibility of using Wikipedia as a resource for annotations of Latin texts. Although Wikipedia is an excellent resource from which to extract many kinds of information (morphological, syntactic and semantic) to be used in NLP tasks on modern languages, it was rarely applied to perform NLP tasks for the Latin language. The work presents the first steps of the development of a POS Tagger based on the Latin version of Wiktionary and a Wikipedia-based semantic annotator.

This paper reports on the development of a rule-based part-of-speech tagger for Classical Tibetan. Far from being an obscure tool of minor utility to scholars, the rule-based tagger is a key component of a larger initiative aimed at radically transforming the practice of Tibetan linguistics through the application of corpus and computational methods.

This paper presents an empirical study on the application of the maximum entropy approach for part-of-speech tagging of Vietnamese text, a language with special characteristics which largely distinguish it from occidental languages. Our best tagger explores and includes useful knowledge sources for tagging Vietnamese text and gives a 93.40%overall accuracy and a 80.69%unknown word accuracy on a test set of the Vietnamese treebank. Our tagger significantly outperforms the tagger that is being used for building the Vietnamese treebank, and as far as we are aware, this is the best tagging result ever published for the Vietnamese language., Nous présentons dans cet article une étude empirique de l'application de l'approche de l'entropie maximale pour l'étiquetage syntaxique de textes vietnamiens. Le vietnamien est une langue qui possède des caractéristiques spéciales qui la distinguent largement des langues occidentales. Notremeilleur étiqueteur explore et inclut des connaissances utiles qui, en terme de performance pour l'étiquetage de textes vietnamiens, fournit un taux de précision globale de 93.40% et de 80.69% pour les mots inconnus sur un ensemble de test du corpus arboré vietnamien. Notre étiqueteur est nettement supérieur à celui qui est en train d'être utilisé pour développer le corpus arboré vietnamien, et à l'heure actuelle c'est le meilleur résultat obtenu pour l'étiquetage de textes vietnamiens.