www.bayesia.com does not fully support your browser (Internet Explorer 6).
We suggest upgrading to IE 7 or downloading Firefox for a more enjoyable web experience.

Studying consumer drivers

We present here how to use BayesiaLab for studying consumer drivers.

The product testing survey is characterized by:

• Baby food tested amongst mothers
• "Look" stage where the mother handles the food before feeding the baby
• "Use" stage where the mother feeds her baby

and we use BayesiaLab to find the drivers of liking.

The described workflow is the following:

• Unsupervised learning to discover all the direct probabilistic relations that hold between the variables
• Variable clustering based on the Kullback-Leibler divergences (non linear measure) corresponding to each arc discovered during unsupervised learning
• Factor induction to create the latent variables corresponding to each identified cluster of variables
• Analysis of each induced Factor to rename it
• Path Analysis by using unsupervised learning restricted to Factor variables and the key variable (Overall Liking here) only for measuring each driver impact
• Unsupervised learning with the Manifest variables to use the Bayesian network for the optimization of the product

Download article (572 kB)