Analysis of consumer data – C13 – Predicting product performance from existing data using machine learning methods – Partial Least Squares (PLS) modelling (On Demand)

Technique to build models to predict one block of correlated data from another. Applications covered include models to predict liking (from sensory/analytic data) or to predict sensory from instrumental/analytic data.

What this course will cover:

• PLS – how it works explained graphically 
• Interpretation and application of key statistics (VIP, Q2)   
• How to build a model selection strategy  
• Application to linking liking and product characteristics (Sensory/ instrumental/ analytic) 
• Issues to consider when predicting liking of future samples 

Pre-requisite: familiarity with PCA (module 3) and regression (module 2) 

This course is not a live session, but an on demand recording you can download instantly, that can then be worked through at your own pace, including workshops with datasets to practice and a certificate of completion at the end. If you wish to experience this as a live session, these are held periodically or you can contact us for a private booking on askqi@qistatistics.co.uk