Simulating and modeling dual market segmentation using PSA framework

Abstract

Market segmentation refers to the analytical process of dividing a broad market into segments taking into account multiple factors such as consumer needs, interests and tastes; it has been considered one of the most important marketing strategies as it helps a business to identify hidden market trends, define target segments, and design marketing plans. Market segmentation may also be viewed as a computational challenge: Given the massive amount of data describing interactions between consumers and commodities, the task is to partition the set of consumers and commodities into subsets that corresponds to market segments—two consumers are in the same segments when they exhibit a similar purchasing pattern, while two products are in the same segments when they are purchased by a similar group of consumers. In this work, we focus on the definition and simulation of market segments. We employ the Propose-Select-Adjust (PSA) framework, introduced in an earlier work [10], to simulate the forming of market segments. Our approach is distributed and can be applied to large and dynamic market data set. The experimental results suggest that the proposed approach is a promising technique for supporting intelligent market segmentation.