Consumer Perception of Red Wine by the Degree of Familiarity Using Consumer-Based Methodology

Capturing and understanding consumers’ perceptions is not a simple quest, particularly for wine, which is one of the most complex beverages. In contrast to the increasing amount of wine import and consumption, studies on how Korean consumers perceive wine characteristics are limited. In this study, two different consumer-based questionnaires, check-all-that-apply (CATA) and rating, were used to compare the discrimination ability of samples and attributes. Consumer data were analyzed and compared to investigate whether the difference in the degree of familiarity with consumption frequency affects wine perception and preference. Consumers discriminated samples and attributes by sample using both scales, CATA and rating. It was confirmed that the CATA citation frequency reflected the rated intensity of the attributes in this study. Consumers who checked or did not check the CATA response rated the intensity of attributes differently. Different consumer subgroups based on familiarity also discriminated the samples effectively. However, users had a higher configuration similarity between the two questionnaires than non-users. Furthermore, the preference for wine might be affected by the degree of familiarity.

Bio-Inspired Techniques in the Clustering of Texts

Today, the development of a large scale access network internet/intranet has increased the amount of textual information available online/offline, where billions of documents have been created. In the last few years, bio inspired techniques which invaded the world of text-mining such, as clustering, represents a critical problem in the digital society especially over the world of information retrieval (IR). The content of this article is a recapitulation of a set of works as a comparative study between the authors' experiments realized by applying a set of bio-inspired techniques (social spiders(SS), 2D Cellular automata (2D-CA), 3D cellular automata (3D-CA), Artificial immune system (AIS), Particle swarm optimization (PSO)) and other techniques founded in literature (Ants Colony Optimization (ACO) and Genetic algorithms (GAs)) for solving the text clustering challenge by using the benchmark Reuter 21785. They analyse the different results in term of entropy, f-measure, execution time, and clusters number in order to find the ideal configuration (similarity measure and text representation method) for each technique. Their objectives are to improve the efficiency of text clustering systems and make decisions that can be the starting point for other researchers.