Compiling a Corpus for Term Extraction Aimed at the Socioterminological Systematization of Mechatronics Terminology

This paper explores the methodological framework for compiling a specialized corpus intended for term extraction, with the objective of supporting the socioterminological systematization of mechatronics terminology. The primary aim of the study is to develop empirically grounded strategies for corpus compilation that enhance the reliability and efficiency of term extraction. It was hypothesized that a corpusdriven approach, when applied exclusively to a didactic subcorpus (mechatronics textbooks), allows for a more precise extraction of term candidates than the approach involving an academic subcorpus (scientific papers in the field of mechatronics). To test this hypothesis, two subcorpora of mechatronics texts written in English were compiled (didactic and academic). Terms were extracted from both the didactic and the academic subcorpus of mechatronics texts and compared to the items extracted by other extraction functions available in Sketch Engine. The comparative statistical analysis has demonstrated that the didactic subcorpus allows for the extraction of more term candidates, while the academic subcorpus produces more noise. Thus, it is argued that a smaller, balanced didactic corpus is more suitable for term extraction aimed at the socioterminological systematization of mechatronics terminology than a larger corpus involving the academic subcorpus. The contribution of the outlined methodology is reflected in greater efficiency in socioterminological systematization of mechatronics terminology. Thus, it is proposed that it may be effectively extended to other interdisciplinary domains.