In the real world, customers usually select products or services based on their assessment of the relevant attributes of the products or services. Moreover, such attribute assessments are often expressed in terms of linguistic data sequences. In order to partition the linguistic data sequences of customers’ assessment on a product or service, we propose a clustering method. In this method, the linguistic data sequences are first converted into fuzzy data sequences and a fuzzy proximity relation is constructed. Then, a fuzzy max-D equivalence relation is derived by using max-D compositions. Based on the fuzzy max-D equivalence relation so obtained, a clustering algorithm is developed and is used to classify the linguistic data sequences into clusters based on which the selection preferences of the customers are identified to develop the customer relationship management (CRM).

clustering, fuzzy proximity relation, max-D transitivity, fuzzy max-D equivalence relation, customer relationship management.