Design And Implementation Of A Parsing-based Machine Translation Algorithm For Numerals From Chinese To English And To German

The thesis first introduced the current situation of the study of machine translation of numerals,then introduced the patterns of numerals in Chinese,English and German.Inspired by X-bar theory in generative grammar,this thesis put forward a syntactic structure to represent the inner structure of Chinese numerals.Based on this structure and the distribution regularities of Chinese coefficient and digit numerals,a set of rules written in context-free grammar which describes the structure of Chinese numerals is designed for Cocke-Younger-Kasami(CYK)parsing algorithm.In this thesis,parsing serves to judge how many Arabic number ‘0's are represented by a Chinese numeral ‘?' if this ‘?' occurs in a specific position in the Chinese numeral to be processed.In order to compensate for the problem of ‘?'-judgement caused by syntactic ambiguity,which is caused by insufficient design of ruleset,a ‘Divide,Conquer and Merge' strategy is applied to realize the automatic machine translation from Chinese numerals to their corresponding English and German numerals.The algorithm is implemented and tested with Python 3.4.4.The final part compares the output of the system developed in this thesis with the outputs of Baidu,Google and Youdao online translations,and proposes 4 aspects where further efforts can be made to improve the performance of the algorithm.The machine translation framework proposed in this thesis is expandable,allowing modules to be added to accomplish translations of numerals between any pair of natural languages.