Year: 2026
Type: dissertation
Abstract: The study explores the potential of large language models (LLMs) for automating key stages of the data analysis lifecycle, including data preprocessing, transformation, visualization, and interpretation. Experimental evaluation of models such as GPT and Claude demonstrates high efficiency in generating code for data cleaning and visualization tasks. However, the reliability of automatically generated insights remains a critical concern due to the risk of hallucinations. The results indicate that LLMs function best as assistive tools within a human-in-the-loop framework, where analysts oversee and validate outputs. Overall, LLMs significantly accelerate routine analytical processes but require careful supervision to ensure accuracy and reliability.
Language: Ukrainian
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7197304037