The paper presents a proposal to identify and analyse consumer purchase behaviour patterns using self-learning GNG neural networks. The employed method is a highly effective tool for analysing multidimensional data sets and e.g. examining purchase patterns. The results suggest that three specific consumer patterns of purchase behaviour exist among the Polish and Slovenian respondents. The findings deepen the knowledge concerning the purchasing behaviour of consumers at the age of 60 +. Parallely, they may be used in shaping international marketing strategies for this consumer group.