Knowledge mining sensory evaluation data: genetic programming, statistical techniques, and swarm optimization

Abstract

Knowledge mining sensory evaluation data is a challenging process due to extreme sparsity of the data, and a large variation in responses from different members (called assessors) of the panel. The main goals of knowledge mining in sensory sciences are understanding the dependency of the perceived liking score on the concentration levels of flavors' ingredients, identifying ingredients that drive liking, segmenting the panel into groups with similar liking preferences and optimizing flavors to maximize liking per group. Our approach employs (1) Genetic programming (symbolic regression) and ensemble methods to generate multiple diverse explanations of assessor liking preferences with confidence information; (2) statistical techniques to extrapolate using the produced ensembles to unobserved regions of the flavor space, and segment the assessors into groups which either have the same propensity to like flavors, or are driven by the same ingredients; and (3) two-objective swarm optimization to identify flavors which are well and consistently liked by a selected segment of assessors.

Development and Biochemical Composition Assessment of Fish Powders from Bangladeshi Indigenous Fish Species and Shelf-Life Characteristics Evaluation During 90 Days of Room Temperature (27°C - 30°C) Storage

“…Among all the powders A. grammepomus, C. punctata and P. puntio samples mostly retained the sensorial qualities and found to be highly acceptable at the end of 90 days storage period. All the samples were found up to the minimum limit of acceptable sensorial recommendation [37] at the end of 90 days room temperature (27˚C -30˚C) storage.…”

“…[36] [37]. Trained panelists (n= 5) were selected from the Department of Nutrition and Food Technology of Jashore University of Science and Technology for the sensory evaluation on the basis of appearance, odor, texture and overall acceptability.…”

A big data approach to computational creativity: The curious case of Chef Watson

“…Given a designed and populated domain knowledge database, a next step is developing a way to generate recipe ideas. Since cuisine naturally has evolutionary properties [29], i.e., cooking styles, techniques, and ingredient choices evolve and even exhibit features like the founder effect, genetic algorithms are one approach to the recipe design problem [30]. Such an approach involves mutating and recombining existing recipes and can produce a myriad of potential recipes.…”

Knowledge Discovery through Symbolic Regression with HeuristicLab

“…The experiment designer simplifies algorithm configuration and makes it very easy to find the best settings for a given problem. Additionally to standard methods for regression HeuristicLab also provides an extensive implementation of symbolic regression based on genetic programming [2], [3]. In this contribution we show the core principle of symbolic regression and discuss how this method can be used for practical data mining applications.…”