Syntactic and Stylistic Analyses of automatically generation subtitles

Abstract

This dissertation addresses the increasing demand for accurate subtitles in video content by conducting grammatical and stylistic analyses of automatic translations, aiming to rectify errors in structure and style that often diminish the viewing experience. The study employs techniques such as part-of-speech tagging, dependency analysis, sentiment analysis, and readability assessment to enhance translation quality, using machine learning and natural language processing on annotated translated texts. The results demonstrate substantial improvements in accuracy, fluency, and user experience, highlighting the value of integrating linguistic knowledge with computational methods in multimedia translation systems, with potential for further research to enhance system robustness