Comparing State-of-the-art Dependency Parsers on the Italian Stanford Dependency Treebank

نویسنده

  • Alberto Lavelli
چکیده

English. In the last decade, many accurate dependency parsers have been made publicly available. It can be difficult for non-experts to select a good off-the-shelf parser among those available. This is even more true when working on languages different from English, because parsers have been tested mainly on English treebanks. Our analysis is focused on Italian and relies on the Italian Stanford Dependency Treebank (ISDT). This work is a contribution to help non-experts understand how difficult it is to apply a specific dependency parser to a new language/treebank and choose a parser that meets their needs. Italiano. Nell’ultimo decennio sono stati resi disponibili molti analizzatori sintattici a dipendenza. Per i non esperti può essere difficile sceglierne uno pronto all’uso tra quelli disponibili. A maggior ragione se si lavora su lingue diverse dall’inglese, perché gli analizzatori sono stati applicati soprattutto su treebank inglesi. La nostra analisi è dedicata all’italiano e si basa sull’Italian Stanford Dependency Treebank (ISDT). Questo articolo è un contributo per aiutare i non esperti a capire quanto è difficile applicare un analizzatore a una nuova lingua/treebank e a scegliere quello più adatto.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Improving a Strong Neural Parser with Conjunction-Specific Features

While dependency parsers reach very high overall accuracy, some dependency relations are much harder than others. In particular, dependency parsers perform poorly in coordination construction (i.e., correctly attaching the conj relation). We extend a state-of-the-art dependency parser with conjunction-specific features, focusing on the similarity between the conjuncts head words. Training the e...

متن کامل

Comparing Italian parsers on a common Treebank: the EVALITA experience

The Evalita ’07 Parsing Task has been the first contest among parsing systems for Italian. It is the first attempt to compare the approaches and the results of the existing parsing systems specific for this language using a common treebank annotated using both a dependency and a constituency-based format. The development data set for this parsing competition was taken from the Turin University ...

متن کامل

Evalita’09 Parsing Task: constituency parsers and the Penn format for Italian

The aim of Evalita Parsing Task is at defining and extending the state of the art for parsing Italian by encouraging the application of existing models and approaches. Therefore, as in the first edition, the Task includes two tracks, i.e. dependency and constituency. This second track is based on a development set in a format, which is an adaptation for Italian of the Penn Treebank format, and ...

متن کامل

Converting Italian Treebanks: Towards an Italian Stanford Dependency Treebank

The paper addresses the challenge of converting MIDT, an existing dependency– based Italian treebank resulting from the harmonization and merging of smaller resources, into the Stanford Dependencies annotation formalism, with the final aim of constructing a standard–compliant resource for the Italian language. Achieved results include a methodology for converting treebank annotations belonging ...

متن کامل

An annotation scheme for Persian based on Autonomous Phrases Theory and Universal Dependencies

A treebank is a corpus with linguistic annotations above the level of the parts of speech. During the first half of the present decade, three treebanks have been developed for Persian either originally or subsequently based on dependency grammar: Persian Treebank (PerTreeBank), Persian Syntactic Dependency Treebank, and Uppsala Persian Dependency Treebank (UPDT). The syntactic analysis of a sen...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2016