Regression Modelling of Key Responses (e.g. Liking) with curvature option
The task: When product developers investigate consumer liking for a group of products, they may also choose to collect sensory data or make analytical measurements at the same time, in the hope of building predictive models that guide development by showing how consumer liking is related to measureable – and therefore changeable – product characteristics.
The problem: The set-up and understanding of the statistical techniques for these types of models can be time consuming and complex, especially when the model is more than just a straight line. Not all stats software easily models curvature in the response.
The solution: The Qi Sensory Drivers Application is a web-based regression tool that allows developers with sensory-consumer data to rapidly target the most important drivers of product liking (or other key response variables) and allows for curvature.
Sensory Drivers Application
Once the sensory and consumer means are loaded into the app, the user identifies the target liking variable, and the software quickly produces a list of drivers sorted by importance.
The software also shows whether each driver acts best as a linear or quadratic predictor of liking, illustrating the best choice with a series of regression plots, making it simple to choose.
Once the model is finalised the results can be exported in an easy to use format (.csv and .png files).
- Web based
- Simple to use
- Easy data file input structure
- Allows for missing data
- Complex stats method made easy
- Fast to calculate the models
- Interactive user assessment of curvature and model choice
- Simple summary statistics presented
- Clear visualisations of models
- Exportable results
$392.34 – $980.85 (Excluding any applicable taxes)