Setence alignment for aparallel corpus: antpconc vs unitex  

Oleh:

Prihantoro

Jenis: Article from Proceeding
Dalam koleksi: KOLITA 15 : Konferensi Linguistik Tahunan Atma Jaya Kelima Belas, page 295-299.
Topik: parallel corpus; AntPConc; XAlign; cross segmental translation; punctuations
Fulltext: 295 Prihantoro - OK.pdf (525.37KB)

Ketersediaan

Perpustakaan PKBB Nomor Panggil: 406 KLA 15 Non-tandon: 1 (dapat dipinjam: 1) Tandon: 1

Reserve Lihat Detail Induk

Isi artikel

Parallel corpus is closely related to machine translation, but can also be used as a reference material for translation studies. In this case, the architecture of a parallel corpus is very significant as even smallest ‘bugs’ may disturb users when the parallel corpus is exploited further either machine translation or translation studies. One of the variables that this paper looks into is the cross segmental translation. This is important as alignment of parallel texts is usually one-on-one (one sentence on the source language is always translated as one sentence in the target language); but in practice, alignment is not necessarily one-on-one. Tools compared in this paper are AntPConc (Anthony, 2014) and Unitex/XAlign (Paumier, 2014). AntConc and Unitex are corpus processing tools known for the processing of a monolingual corpus. Unlike XAlign, which is an integrated functionality within Unitex, AntPConc is a separate tool from AntConc; therefore, some of the functionalities in AntConc might be absent in AntPConc. In relation to the architecture of the parallel corpus, this paper describes strengths of the two tools, and functionalities that they might improve.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali