The use of context-free grammar in support of Slovak-English rule-based machine translation.

摘要

Machine translation has been developing for many decades. The rule-based machine translation approach has been implemented in computational linguistics since its very early days. Human involvement in this approach is significant as it is the human agent who creates the rules so that the system generates desirable outcomes. The benefit of this approach is that it analyzes the input on the syntactic and - to an extent - semantic levels. The downside to rule-based machine translation is that it requires a deep linguistics knowledge as well as long time to prepare the rules.;The goal of this paper is to focus on Slovak grammar and its morphological and syntactic complexities. Once identified, possible solutions that will help manage these complexities will be proposed. The goal of this thesis is to build an extensive set of rules the combination of which will result in proper English translation of the sentences in my training dataset. The rules will entail morphological, syntactic parsing, transfer as well as some post-processing rules that are all necessary for executing translation. Context-free grammar and its rules will be utilized in preparation of the grammar that will serve as a basis for parsing and transfer of Slovak input into English output.