Multivariate and Probabilistic Analyses of Sensory Science Problems

Jean-François Meullenet, Ph.D., is associate professor of Sensory Science in the Department of Food Science at the University of Arkansas, Fayetteville, AR. Dr. Meullenet conducts research in the area of sensory science and his expertise encompasses sensory and consumer science, rheology and modeling of food perception. Rui Xiong, Ph.D., is a research scientist with the Consumer Science Insights, Unilever Home & Personal Care, Trumbull, CT, USA. Christopher J. Findlay, Ph.D., is president of Compusense, Inc., Guelph, Ontario, Canada. He is associate editor for sensory evaluation for Food Research International.

• Introduction 3Chapter 1 A Description of Sample Data Sets Used in Further Chapters 91.1 A Description of Example Data Sets 9References 25Chapter 2 Panelist and Panel Performance: A Multivariate Experience 272.1 The Multivariate Nature of Sensory Evaluation 272.2 Univariate Approaches to Panelist Assessment 292.3 Multivariate Techniques for Panelist Performance 322.4 Panel Evaluation through Multivariate Techniques 432.5 Conclusions 46References 47Chapter 3 A Nontechnical Description of Preference Mapping 493.1 Introduction 493.2 Internal Preference Mapping 493.3 External Preference Mapping (PREFMAP) 583.4 Conclusions 66References 67Chapter 4 Deterministic Extensions to Preference Mapping Techniques 694.1 Introduction 694.2 Application and Models Available 694.3 Conclusions 89References 94Chapter 5 Multidimensional Scaling and Unfolding and the Application of Probabilistic Unfolding to Model Preference Data 955.1 Introduction 955.2 Multidimensional Scaling (MDS) and Unfolding 965.3 Probabilistic Approach to Unfolding and Identifying the Drivers of Liking 985.4 Examples 100References 109Chapter 6 Consumer Segmentation Techniques 1116.1 Introduction 1116.2 Methods Available 1116.3 Segmentation Methods Using Hierarchical Cluster Analysis 113References 126Chapter 7 Ordinal Logistic Regression Models in Consumer Research 1297.1 Introduction 1297.2 Limitations of Ordinary Least Square Regression 1297.3 Odds Odds Ratio and Logit 1307.4 Binary Logistic Regression 1337.5 Ordinal Logistic Regression Models 1447.6 Proportional Odds Model (POM) 1447.7 Conclusions 160References 160Chapter 8 Risk Assessment in Sensory and Consumer Science 1638.1 Introduction 1638.2 Concepts of Quantitative Risk Assessment 1648.3 A Case Study: Cheese Sticks Appetizers 1668.4 Conclusions 176References 176Chapter 9 Application of MARS to Preference Mapping 1799.1 Introduction 1799.2 MARS Basics 1799.3 Setting Control Parameters and Refining Models 1879.4 Example of Application of MARS 1889.5 A Comparison with PLS Regression 201References 205Chapter 10 Analysis of Just About Right Data 20710.1 Introduction 20710.2 Basics of Penalty Analysis 20810.3 Boot Strapping Penalty Analysis 21010.4 Use of MARS to Model JAR Data 21210.5 A Proportional Odds/Hazards Approach to Diagnostic Data Analysis 21510.6 Use of Dummy Variables to Model JAR Data 220References 233Index 237

?This technical work provides a useful insight into the solution of a number of pertinent problems in sensory science. This is an excellent work for sensory specialists and challenges the reader to consider alternate strategies for handling sensory data.? ( Journal of Dairy Technology, May 2009)