Intelligent Knowledge Of The Machine Translation System Auxiliary To Obtain

Machine Translation needs a huge knowledge base as well as the translation mechanism and software environment using the knowledge to translate. The establishment and maintenance of knowledge base are very hard and difficult. Therefore, knowledge acquisition through multiple channels is one of the most important and difficult problems in MT.In this treatise, based on the theory of Intelligent Machine Translation, knowledge acquisition method based on morphemics, corpus and feedback knowledge is proposed. An integrated knowledge acquisition environment, on the basis of IMT E/C, including these three methods above, is designed and implemented. Knowledge acquisition based on morphemics solves the acquisition problem of the attributes of derivative words and helps to maintain the consistency of the lexicon. Knowledge acquisition tool based on corpus can analyze automatically 50% of the sentences. The verb information thus acquired has small granite, good conformity and are objective. So it helps to improve the efficiency of the acquisition of verb entry. Knowledge acquisition tool based on feedback information allows user to correct the wrong translation. At the same time, it infers backwards according to the user operation to find the errors or incompleteness in lexicon and rule base, even give some suggestion for revision. In addition, more convenient editor for lexicon is included in the integrated environment, which makes the input of lexicon entries easier and more visual.This integrated mechanism, based on the original acquisition method, using the achievement of current MT knowledge acquisition and the experience of the original method, is oriented directly to the current requirement of IMT E/C. It not only drastically shortens the period of...