National corpus and lexical infrastructure: the experıence of building the Azerbaijani language lexical database

This article presents the methodology and outcomes of a multi-stage research project aimed at building the first comprehensive lexical database of the Azerbaijani language, developed within the framework of corpus linguistics and statistical lexicography.The database functions as a foundational linguistic module within the national corpus, providing structured lexical inventories for both academic research and technological applications.During the research, a corpus of 520 million word-forms from various functional styles was collected, cleaned, and structured using specialized software.As a result of automatic processing, 2, 918, 910 word-forms were initially identified; through several subsequent stages of technical and lexical filtering, the database was refined.Ultimately, 175, 521 lexical units were compiled into structured word lists, sorted by frequency and in alphabetical order.Additionally, 93, 287 words not found in the orthographic dictionary of Azerbaijani language were identified and presented in a separate list.The article elaborates on the structure of the national corpus, its components (fiction, publisistic, scientific, official, spoken, and educational texts), concordances, and linguistic analyzers.The lexical database is presented as a strategic resource for both theoretical research and practical applications such as natural language processing, automatic translation, text-to-speech systems, and artificial intelligence models.The study demonstrates significant scientific and practical outcomes in the development of digital lexical resources for the Azerbaijani language.