Introducing Two Criteria for Future Data Prediction with ML Models

Learn about unbiased machine learning model performance assessment in this conference talk from Data Science Conference Europe 2023. Explore how training databases directly influence model outcomes and understand the scientific community's use of standardized databases for result comparability. Delve into the challenges of introducing new databases and modifying data preparation processes, while discovering novel evaluation criteria that connect database characteristics with machine learning metrics. Through a practical example focused on predicting subsequent data points, gain valuable insights into transparent evaluation methodologies for machine learning models in today's rapidly evolving AI landscape. Originally presented by Dusan Radivojevic during the Tech Tutorials stream, this comprehensive discussion equips data scientists and AI practitioners with essential knowledge for implementing more accurate and transparent model evaluation practices.