The expectation-maximization (EM) algorithm for incomplete data with highly diverse missing data patterns can be computationally expensive. A partial expectation-maximization (PEM) algorithm is developed to ease this computational burden. This PEM algorithm circumvents the need for a traditional E-step by performing a partial E-step that reduces the Kullback-Leibler divergence between the conditional distribution of the missing data and the distribution of the missing data given the observed data. The PEM and EM algorithms are compared in terms of computation time and convergence on simulated data. The PEM algorithm is illustrated using a latent Gaussian mixture model to cluster a white bread sensory analysis dataset.

As technology, economies, and populations change, so too will the sensory methods, tools, and approaches used to assess difference. Discrimination testing will continue to see changes with developments in technology, particularly the application of handheld and wearable devices, advancements in the understanding of psychophysics and statistics, a more philosophical approach to the criterion used to make difference judgments, and the impact of climate change and the growth of the planet's population exerting pressure on the supply of raw materials. The future of discrimination testing is faster, more accurate, and more oriented to differences that are meaningful to consumers.

In sensory evaluation, it may be necessary to design experiments that yield incomplete data sets. As such, sensory scientists will need to utilize statistical methods capable of handling data sets with missing values. This article demonstrates the advantages of a model-based imputation procedure that simultaneously accounts for heterogeneity while imputing. We compare this model-based approach to the current state-of-the-art imputation procedures using two real data sets that arose from central location tests. These data sets contain missing values by design. In addition, these data sets have two data sets nested within each of them. We use these nested data sets to validate the results. Compared to the considered state-of-the-art imputation procedures, we find evidence that the model-based approach is able to recover the group structure and key characteristics of the data sets when a high percentage of the data are missing.Practical applicationsThe model-based imputation procedure presented in this manuscript is used to analyse incomplete multivariate data sets arising from central location tests. It can be used to analyse incomplete mulitvariate data sets where there is correlation among the variables. Examples include data arising from high fatigue studies, just-about-right scale, free choice profiling, or from the ideal profile method.

Liking studies are designed to ascertain consumers likes and dislikes on a variety of products. However, it can be undesirable to construct liking studies where each panelist evaluates every target product. In such cases, an incomplete-block design, where each panelist evaluates only a subset of the target products, can be used. These incomplete blocks are often balanced, so that all pairs occur the same number of times. While desirable in many situations, balanced incomplete blocks have the disadvantage that, by their nature, they cannot favor placing dissimilar products next to one another. A novel incomplete-block design is introduced that utilizes the target product’s sensory profile to allocate products to each panelist so that they are, in general, as dissimilar as possible while also ensuring position balance. The resulting design is called a sensory informed design (SID). Herein, details on the formulation of SIDs are given. Data arising from these SIDs are analyzed using a simultaneous clustering and imputation approach, and the results are discussed.

Clustering consumer data reveals important information to help refine products for specific market segments. There are compelling reasons to use incomplete block designs to collect consumer data; however, this presents the challenge of dealing with missing data. The purpose of this workshop is to investigate the effect of different imputation techniques on the results of cluster analysis of balancedincomplete-block (BIB) data. BIB designs and possible ways to incorporate information from sensory data into the design (cf. Browne et al., 2013) will be reviewed. Then, commonly used imputation and clustering approaches will be discussed. After this background material has been covered, expectationmaximization (EM) algorithm-based imputation will be presented. The remainder of the workshop will be interactive and hands on. Participants will be given two data sets (provided by Compusense Inc.) resulting from sensory informed BIB designs. Using R software, which will also be provided, participants will be invited to use EM algorithm-based imputation and compare the results to other approaches. Reference:

The design of new products for consumer markets has undergone a major transformation over the last 50 years. Traditionally, inventors would create a new product that they thought might address a perceived need of consumers. Such products tended to be developed to meet the inventors own perception and not necessarily that of consumers. The social consequence of a top-down approach to product development has been a large failure rate in new product introduction. By surveying potential customers, a refined target is created that guides developers and reduces the failure rate. Today, however, the proliferation of products and the emergence of consumer choice has resulted in the identification of segments within the market. Understanding your target market typically involves conducting a product category assessment, where 12 to 30 commercial products are tested with consumers to create a preference map. Every consumer gets to test every product in a complete-block design; however, many classes of products do not lend themselves to such approaches because only a few samples can be evaluated before `fatigue' sets in. We consider an analysis of incomplete balanced-incomplete-block data on 12 different types of white bread. A latent Gaussian mixture model is used for this analysis, with a partial expectation-maximization (PEM) algorithm developed for parameter estimation. This PEM algorithm circumvents the need for a traditional E-step, by performing a partial E-step that reduces the Kullback-Leibler divergence between the conditional distribution of the missing data and the distribution of the missing data given the observed data. The results of the white bread analysis are discussed and some mathematical details are given in an appendix.

International Journal of Food Science & TechnologyVolume 48, Issue 1 p. 215-219 Short communication Experimental consideration for the use of check-all-that-apply questions to describe the sensory properties of orange juices Youngseung Lee, Youngseung Lee Department of Food Science and Nutrition, Dankook University, Jukjeon-Dong 448–701, Yongin, Gyeonggi-Do, KoreaSearch for more papers by this authorChris Findlay, Chris Findlay Compusense Inc., 679 Southgate Drive, Guelph, Ontario, Canada, N1G 4S2Search for more papers by this authorJean-François Meullenet, Corresponding Author Jean-François Meullenet Department of Food Science, University of Arkansas, 2650 N. Young Avenue, Fayetteville, AR, 72704 USA Correspondent: Fax: +(479) 575 6936; e-mail: [email protected]Search for more papers by this author Youngseung Lee, Youngseung Lee Department of Food Science and Nutrition, Dankook University, Jukjeon-Dong 448–701, Yongin, Gyeonggi-Do, KoreaSearch for more papers by this authorChris Findlay, Chris Findlay Compusense Inc., 679 Southgate Drive, Guelph, Ontario, Canada, N1G 4S2Search for more papers by this authorJean-François Meullenet, Corresponding Author Jean-François Meullenet Department of Food Science, University of Arkansas, 2650 N. Young Avenue, Fayetteville, AR, 72704 USA Correspondent: Fax: +(479) 575 6936; e-mail: [email protected]Search for more papers by this author First published: 11 September 2012 https://doi.org/10.1111/j.1365-2621.2012.03165.xCitations: 40Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL References Ares, G., Barreiro, C., Deliza, R., Gimenez, A. & Gambaro, A. (2010). Application of a check-all-that-apply question to the development of chocolate milk desserts. Journal of Sensory Studies, 25, 67– 86. Dillman, D.A., Smyth, J.D., Christian, L.M. & Stern, M.J. (2003). Multiple Answer Questions in Self-Administered Surveys: The Use of Check-all-that-Apply and Forced-Choice Question Formats. San Francisco, CA: Meetings of the American Statistical Association, August. Dooley, L., Lee, Y.S. & Meullenet, J.F. (2010). The application of check-all-that-apply (CATA) consumer profiling to preference mapping of vanilla ice cream and its comparison to classical external preference mapping. Food Quality and Preference, 21, 394– 401. Hottenstein, A.W., Taylor, R. & Carr, B.T. (2008). Preference segments: a deeper understanding of consumer acceptance or a serving order? Food Quality and Preference, 19, 711– 718. Israel, G.D. & Taylor, C.L. (1990). Can response order bias evaluations? Evaluation and Program Planning, 13, 365– 371. Krosnick, J.A. (1991). Response strategies for coping with the cognitive demands of attitude measures in surveys. Applied Cognitive Psychology, 5, 213– 236. Krosnick, J.A. (1999). Survey research. Annual Review of Psychology, 50, 537– 567. Krosnick, J.A. & Alwin, D.F. (1987). An evaluation of a cognitive theory of response-order effects in survey measurement. Public Opinion Quarterly, 51, 201– 219. Lado, J., Vicente, E., Manzzioni, A. & Ares, G. (2010). Application of a check-all-that-apply question for the evaluation of strawberry cultivars from a breeding program. Journal of the Science of Food and Agriculture, 90, 2268– 2275. Lee, Y.S. & Meullenet, J.F. (2010). Comparison of eliminating first order samples for minimizing first serving order bias to data corrections. Food Science and Biotechnology, 19, 703– 709. Parente, M.E., Manzoni, A.V. & Ares, G. (2011). External preference mapping of commercial antiaging creams based on consumers' responses to a check-all-that-apply question. Journal of Sensory Studies, 26, 158– 166. Wakeling, I.N. & MacFie, H.J.H. (1995). Designing consumer trials balanced for first and higher orders of carry-over effect when only a subset of k sample from t may be tested. Food Quality and Preference, 6, 299– 308. Citing Literature Volume48, Issue1January 2013Pages 215-219 ReferencesRelatedInformation

Dr Findlay spoke on developments in consumer research at the SAAFoST Northern branch meeting, attended by members of SAAFoST and students from the University of Pretoria.

Statistical issues are discussed within the context of understanding the practical considerations that apply to quality control. The concepts of risk and power are presented and related to specific test types. Worked examples are provided for four of the most common sensory quality control procedures.

Tienda online donde Comprar Multivariate and Probabilistic Analyses of Sensory Science Problems al precio 175,77 € de Jean-Francois Meullenet | Rui Xiong | Christopher Findlay, tienda de Libros de Medicina, Libros de Quimica - Quimica

This chapter contains sections titled: Introduction Concepts of QRA A Case Study: Cheese Sticks Appetizers Conclusions

Training targets were established using descriptive analysis profiles of 20 commercial red wines produced by a well-trained, experienced determination panel. After recruitment, screening and a basic sensory orientation of ten 2h common training sessions, 16 inexperienced panelists were divided by lottery into two panels. The control panel received a more conventional performance debriefing at the end of each training session. The experimental panel only received immediate graphical computerized feedback while in sensory booths. Both panels evaluated the same 20 wines and used the same scales and attributes. Panels were calibrated and responses compared to training targets. Performance was monitored daily as panels continued over a three-week period. Distance from target measurements showed similar improvement trends for both groups as measured by panelist and panel calibration. Results suggest the effectiveness of the feedback calibration method (FCM) in providing unbiased and effective training.

The growth of Internet use and subsequent uptake of e-commerce have significantly changed the business environment. Automation of procurement processes and functions has magnified opportunities and exacerbated the inherent risks in public sector procurement. The Scottish Executive has established a common e-procurement approach for public procurement agencies in Scotland. This paper examines the Scottish model and compares it with the global e- procurement experience to reveal the benefits and dangers of centralised requirements and aggregated orders. The potential impact of e-procurement policy on local Small and Medium Enterprises (SMEs) is analysed and recommendations made for future policy. Case studies demonstrating benefits to Scottish SMEs are reviewed. The paper concludes that the efficiency benefits of e-procurement make its widespread adoption inevitable and while e-procurement presents opportunities for SMEs, public policy makers should be alert to the risks of aggregation, centralisation and skills gaps.

Computer technology is changing rapidly as is the scope and use of the Internet. These tools are being applied to a broad range of quality control activities, including sensory evaluation. The main areas of impact of this technology are in test design, collection of data, tabulation, storage, statistical analysis and reporting of the data in real time over great distances. Effective quality systems can be constructed using anything from the simplest spreadsheet programs through to sophisticated integrated quality control systems operating over corporate networks. This article provides an overview of the tools that are available and discusses a specific case as an example of a starting point for computerizing sensory quality control.