POLENG–Adjusting a Rule-Based Polish–English Machine Translation System by Means of Corpus Analysis
Krzysztof Jassem, Filip Graliński, Grzegorz Krynicki · 2000
Introduction POLENG is a transfer text-to-text Polish-English MT system. It operates on an electronic dictionary annotated for morphological, syntactic, and partly semantic information. The dictionary is based on the corpus of Polish texts from the domain of computer science. The local version of the system has the form of a two-window editor. The client-server version is available at http://poleng.wmid.amu.edu.pl/. The system is intended to deal with HTML Polish texts as well as cooperate with Microsoft Word. The detailed structure of the system is described in (Wypych 2000). In our short presentation we will concentrate on corpus-based methods we adopt in order to improve the efficiency of the translation. 2. General properties of the POLENG system 1.1 Translation algorithm The POLENG system is based on a uni-directional translation algorithm. The tree for representing English output expressions is built in the course of the analysis of Polish input.