An unsupervised learning algorithm to store pattern recognition

One of the major challenges of any commercial company is to be able to manage the whole information available from their customers and turn it into useful knowledge to support successful business decisions. Moreover, usually that information comes from different sources with the added problems of subjectivity and ambiguity, especially as far as qualitative data is concerned. In this paper, an unsupervised learning algorithm is applied for classifying the customer's behavior (retailer's behavior) of a Spanish supplier. The goal of the classification is to recognize whether there are both existing and latent separate types of customers.