pipette
ENEnglish

DagFlow AI: Agentic Synthetic Data Generation for Personalized Statistics Learning

David Arthur, Ranjini Grove, Zaher Kmail, Curtis Atkisson

Preprint

En palabras de los autores

To provide personalized learning experiences in courses that teach data analysis skills, instructors often search for datasets that appeal to a student's interests, align with specific learning objectives, and are appropriate for the student's skill level. Datasets that match all three criteria can be difficult to find. DagFlow-AI is an agentic AI data generation tool that creates realistic, synthetic datasets tailored to these criteria. Rather than mimicking existing datasets or generating observations directly, DagFlow-AI uses the broad knowledge and reasoning capabilities of large language models (LLMs) to specify the components of transparent, reproducible data-generating mechanisms from which observations are generated programmatically. We illustrate how the tool can be used to construct a realistic, synthetic dataset and describe how it aligns with both the Guidelines for Assessment and Instruction of Statistics Education (GAISE) and the Problem, Plan, Data, Analysis, and Conclusion (PPDAC) framework for teaching data analysis

Resultado principalEl resumen no menciona limitaciones.

Apareció: viernes, 23 de octubre. arXiv. Preprint, todavía sin revisión por pares.

Comentario de los autores: 19 pages, submitted to Technological Innovation in Statistics Education