Clustering around latent variables approach to detect panel disagreement in three–way tea sensory evaluation

Sensory evaluation plays a remarkable role in maintaining the quality standards of beverages such as tea. The quality, which determines the price of tea, is evaluated by professional tea tasters. Uncertainty and vagueness of sensory evaluation has been a serious issue in selection of good quality tea. An issue existing when analyzing sensory data to detect panel disagreement is that data of three dimensional (three–way) or higher are often reduced to two–way data. Present study aimed to investigate the possibility of using Clustering around Latent Variables for three–way data (CLV3W) method to detect panel disagreement in sensory data of Sri Lankan tea. A three–way data set, 8 tea tasters× 13 tea growing regions×6 attributes, for each month with four replicates (4 factories per region), for a period of one calendar year were used for the study. When CLV3W analysis was performed separately for each month data, it was found two–clusters (two–latent components) exist for the data. Attribute loadings of colour and strength indicated that they are represented by the first latent component, and those loadings for brightness, flavour, aroma, and quality indicated that they are represented by the second component. Region scores for the two components revealed grouping of regions indicating some regions have similar tea quality attributes. Assessor weights varied across latent components and months. However, by examining assessor weights it was possible to identify assessors those who were in agreement and those who were in disagreement for each component. This finding demonstrated the fact that certain assessors are more sensitive to certain attributes and thus in order to detect tea with certain attributes, appropriate assessors can be employed. Therefore, CLV3W method is a useful method to detect disagreement between assessors, especially in 3–way data, and it can effectively be used when selecting assessors for sensory evaluation.

[1]  Michel Tenenhaus,et al.  Measurement of the reliability of sensory panel performances , 2006 .

[2]  A.Y.S.L Silva,et al.  A Study to Recruit and Train the Product oriented Sensory Panel , 2014 .

[3]  Pascal Schlich,et al.  Defining and Validating Assessor Compromises About Product Distances and Attribute Correlations , 1996 .

[4]  X. Zhu,et al.  Correlation between Sensory Evaluation Scores and Flavor Attributes of Black Teas from Around the World , 2018 .

[5]  M. Ranatunga,et al.  Biochemical Characteristics of Tea (Camellia L. spp.) Germplasm Accessions in Sri Lanka: Correlation between Black Tea Quality Parameters and Organoleptic Evaluation , 2014 .

[6]  Tormod Næs,et al.  Measuring validity in sensory analysis , 1995 .

[7]  Luis Guerrero,et al.  A cross-national consumer segmentation based on food benefits: The link with consumption situations and food perceptions , 2012 .

[8]  Kenneth Tze Kin Teo,et al.  A review on intelligent sensory modelling , 2016 .

[9]  E. Vigneau,et al.  Clustering of Variables Around Latent Components , 2003 .

[10]  Tom F. Wilderjans,et al.  CLV3W: A clustering around latent variables approach to detect panel disagreement in three-way conventional sensory profiling data , 2016 .

[11]  N. Guzman,et al.  Comparison of Results from Cupping and Descriptive Sensory Analysis of Colombian Brewed Coffee: Coffee: Comparing Cupping and Descriptive Analysis , 2014 .

[12]  El Mostafa Qannari,et al.  A method for panelists’ consistency assessment in sensory evaluations based on the Cronbach’s alpha coefficient , 2014 .

[13]  C. Walgampaya,et al.  Use of Mandel’s bundle of lines model to improve agreement of a panel of tea tasters , 2018 .

[15]  Rosaria Romano,et al.  Monitoring panel performance within and between sensory experiments by multi-way analysis , 2011 .

[16]  Rosaria Romano,et al.  Combining analysis of variance and three‐way factor analysis methods for studying additive and multiplicative effects in sensory panel data , 2015 .

[17]  Tormod Næs,et al.  Performance indices in descriptive sensory analysis – A complimentary screening tool for assessor and panel performance , 2013 .

[18]  Physico-chemical characterization and fuzzy logic modeling of sensory evaluation for market Ghewar , 2017 .

[19]  L. Carbonell,et al.  Sensory analysis of Spanish mandarin juices. Selection of attributes and panel performance , 2007 .

[20]  U. Pankiewicz,et al.  Sensory and chemical evaluation of tea brews prepared with the assistance of ultrasound , 2019, Ciência Rural.

[21]  Tom F. Wilderjans,et al.  Clusterwise Parafac to identify heterogeneity in three-way data , 2013 .

[22]  Per B. Brockhoff,et al.  Monitoring panel performances with the Mixed Assessor Model. Meta-analysis of the SensoBase , 2012 .