Fuzzy Clustering in Travel and Tourism Analytics

Cluster analysis is one of the most common techniques used in tourism and marketing research to find homogeneous groups of units (i.e. tourists or consumers) according to a set of segmentation variables. Since it may be too strict to assume that each unit belongs to one cluster 100%, fuzzy clustering algorithms have been suggested in the market segmentation literature, which allow the same unit to be assigned to more than one cluster with a membership degree. Often, individual judgements, like emotions, motivations, or satisfactions, are captured making use of qualitative scales, such as Likert-type scales. Information collected using such scales are vague and uncertain mainly because respondents have to convert their judgements on a linguistic expression, often expressed in numerical values, the meaning of which can be different per respondent. To reduce this source of vagueness, ordinal variables can be transformed into fuzzy variables before the adoption of a clustering algorithm. This operation requires the modification of traditional fuzzy clustering algorithms in order to cope with the fuzzy nature of the segmentation variables. Two kinds of fuzzy clustering algorithms for fuzzy data are theoretically presented in this chapter, together with two empirical case studies.