الفهرس 
التَّرجمة الآليَّة ومواردُها: التَّعريفُ بها وتطوُّرُهايوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 308 |  |  |
| | | from | 308 | | |
المستخلص
The study aims to build and develop a methodology to benefit from machine translation resources; To build algorithms for language processing at the level of analysis and generation; With the aim of developing applications for machine translation between Arabic and English, and improving its output, by building a parallel corpus between the two languages, which helps in extracting a set of patterns, structures and rules that states the problems of machine translation, and contribute to improving its outputs with better accuracy and higher quality by using automatic analysis tools (concordances tools, morphological analyzers, semantic analyzers, taggers, wordnet, contextual concordances) and other automatic analysis tools for natural languages.
The research was titled ”Machine translation resources between Arabic and English... Computational linguistic processing”; where the research intended Contemporary Standard Arabic to study, trying to build a linguistic resource that contributes to improving the output of machine translation between Arabic and English, using linguistic corpus as a means of collecting and storing linguistic material to allow using automatic analysis tools to deal with it; To extract information that is hoped to be used in building the desired resource, and to keep the progress in computational linguistics.
The research took the language of film subtitles between Arabic and English as its subject matter, because it is the closest to the reality of the language, and the daily use of it; where the material was collected in the light of the theory of statistical samples, using the descriptive approach in dealing with the linguistic texts collected; where it is useful in describing and observing its features without editing, except with regard to linguistic review and correction of errors.
The research used a set of linguistic analysis tools; To extract the information hoped to be presented by the study sample; where concordances are used to index the raw data, extract the occurrence information, and use the morphological analyzers to find the morphological features of the corpus vocabulary, as well as using taggers; To annotate the vocabulary of raw texts with the most prominent linguistic features such as parts of speech information, and also use contextual concordances to track the different contexts of vocabulary and phrases within the corpus in Arabic and English, in addition to using the BERT system to extract the study sample that expresses machine translation problems between the two languages.
The research concluded that the lack of linguistic resources in the Arabic language hinders its progress in the field of machine translation, especially as it is a language that has a somewhat complex morphological and syntactic system in contrast to the English language, confirming the need to take advantage of the already existing resources such as the “WordNet” and “FrameNet”, and working on developing them to support the Arabic language in the field of machine translation, as well as striving to build more various resources, whether linguistic, computational or lexical resources; To support our Arabic language.
The research was able to build a parallel corpus between two languages, showing the linguistic features that control the phrases and sentences in the two languages; To take advantage of the similarities and differences in improving machine translation output; where a group of patterns and structures that suffer from poor processing were extracted, which led to translations with low accuracy and poor quality; Take advantage of these sentences and patterns in suggesting solutions to machine translation problems and improving its performance.