As regulatory acceptance of models for decision-making grows, so too does the need for optimally fit models, free as possible from human bias. This places high demands on the modeling workbench, which must build and fine-tuning thousands of models across a vast search space, then select a winner according to standard goodness-of-fit and custom criteria.
Join us as we outline the components required of today’s popPK modeling workbench, from tools for model construction and selection to diagnostics reporting and change logs. We’ll dive deep into the importance for selection algorithms that go beyond exhaustive search to include genetic, Gaussian, and other approaches. Finally, we’ll showcase how these capabilities come together in Pirana, Certara’s popPK modeling workbench.
Webinar agenda:
• Exploration of machine learning model selection
• Demonstration of model evaluation and reporting tools for both machine learning and manual model runs
• Brief overview of Pirana
• Q&A with experts Mark Sale and James Craig
