by Katya Vladislavleva, Chief Data Scientist@ Evolved Analytics
Sensory evaluation science is hard. Experiments are expensive, and it is difficult to collect enough data. Measurements have too many factors with unknown cause and effect relationships. Their co-dependencies are often non-linear and validation of the dependencies requires considerable time and research effort. This creates high demand for data-driven solutions to efficiently extract actionable knowledge from sensory evaluation data and help focus research. This talk focuses on the impact of using non-linear data-driven modeling technology to effectively extract taste and flavor relationships from experimental data, design better experimental protocols, predict consumer preferences, identify ingredients that drive liking and ultimately design flavors with given desirability profiles.
We demonstrate the recent technological advances in the area of advanced predictive analytics and data science on two case studies.