ModelSelect Data and Results

Despite the rapid growth of AI, selecting suitable models poses a challenge, especially for Research Software Engineers (RSEs) who often lack deep expertise in the field. This study addresses this challenge by proposing a structured decision model to guide RSEs in choosing AI models tailored to their needs.

The proposed framework integrates multi-criteria decision-making (MCDM) principles and combines objective performance metrics with subjective user sentiment analysis. By collecting data from public sources like documentation, repositories, and community feedback, the model facilitates trade-off analysis, optimizing for constraints such as computational efficiency and long-term maintainability. Using Design Science Research (DSR) methodology, the model was developed and validated through real-world case studies, demonstrating its effectiveness in enhancing decision-making accuracy.