Frano Mihaljevic, MS: No financial relationships to disclose
Description of session (include background & scientific importance): Structural model selection remains a largely manual and time-consuming process in pharmacometrics, relying on expert judgment and iterative refinement. While a few automation tools exist, they have not seen widespread adoption, often due to limitations in usability, flexibility, or performance. Meanwhile, the field of model-building automation is evolving rapidly, with new approaches and advancements continually emerging. As pharmacometric workflows become increasingly complex, there is a growing need for robust and efficient automated solutions to support model-based drug development.
This session will present a comprehensive comparison of algorithms for automated structural model building, assessing their effectiveness in different pharmacometric scenarios.
By addressing the challenges and opportunities in model building automation, this session aims to provide practical insights into how pharmacometricians can transition from manual selection to intelligent, automated approaches that improve workflow efficiency and model quality.
Learning Objectives:
Upon completion, participant will be able to
• Understand the goals of automated structural model building and recognize the trade-offs involved in algorithmic model building approaches,
• Identify which algorithms perform best in specific scenarios based on empirical comparisons,
• Learn how the new tool can navigate search spaces of different sizes to find optimal models,
• Evaluate how this approach can streamline workflows and enhance model-building outcomes.
