D. E. Chikrin, A. V. Puga · INFORMACIONNYE TEHNOLOGII 2026 · 2026
DOI: 10.17587/it.32.507-518
Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).
The article proposes a component-based approach to assessing the applicability of large language models in decision support systems. The necessity of multidimensional evaluation is substantiated, and an original methodology for component-based assessment is presented, taking into account both the content of the text and its suitability for specific applied contexts. Unlike existing review scales, the approach suggests the use of operational applicability profiles of the response and a decision-making procedure at the level of a specific output. The components capture verifiable properties of the response and serve as the basis for reproducible metrics and evaluation criteria. The approach is oriented toward typical decision support system tasks and can be applied in both expert and automated scenarios. The methodology can be used to verify generated results and enhance the substantiation of trust in them.
No comments yet — start the discussion below.