A Corpus-Based Study on the Consistency of Business English Terminology Translation

Abstract

Accurate semantic representation and terminology consistency are essential for intelligent information processing and cross-domain communication in data-driven engineering systems. To address inconsistencies in business English terminology translation, this study proposes a corpus-based analytical framework that integrates parallel corpus construction, statistical feature extraction, contextual analysis, and quantitative consistency evaluation. A bilingual corpus covering multiple business subdomains is established to investigate the distribution characteristics of terminology translation through lexical overlap analysis, frequency-based screening, expert verification, and entropydriven assessment. By combining translation frequency statistics, dominant translation ratios, Shannon–Wiener diversity indices, and cross-domain distribution entropy, the proposed framework systematically reveals registerdependent translation patterns and semantic constraints underlying terminology variation. Experimental analyses demonstrate that terminology consistency is strongly influenced by document type and contextual characteristics, with legal texts exhibiting significantly higher stability than marketing-oriented documents. The proposed methodology provides an effective data-driven solution for semantic standardization, knowledge organization, and adaptive information management. Beyond translation studies, the framework offers methodological references for intelligent semantic alignment, information fusion, and communication-oriented data processing in engineering systems, supporting potential applications in Electromagnetic Waves, Antennas and Propagation where accurate knowledge representation, adaptive information exchange, and heterogeneous data integration are critical.