Market Segmentation of Innovative Products Using Genetic Algoritms

One way to increase the company's competitiveness is to find new market niches. The market niche is the result of innovations that stimulate hidden, potential demand, as a result of which the company, developing a new market, avoids intense competition and receives a higher rate of return. It is proved that the growing number and complexity of tasks in the field of marketing research, working with a large amount of information, leads to the need to group data. The aim of the study is to develop a universal approach to solving the problem of market segmentation of innovative products based on a combination of genetic algorithm with traditional clustering methods. An ideal market niche can be defined as a compact and isolated series of points, which in some space of characteristics are objects or data elements. The selection of a market niche in the medical equipment market is carried out using a top-down approach. This approach implies the traditional segmentation of customers, which is carried out in the following order: segmentation, segment selection, positioning. It is believed that segmentation is the starting point for the formation of a market niche. To segment the medical equipment market, it is proposed to use cluster analysis methods. According to the results of the analysis, it can be seen that the market segments of potential consumers of medical equipment and consumables of Siemens in Ukraine are characterized by a fairly dense grouping of images of consumers around the center of its cluster in the space of features. The presented genetic clustering algorithm is flexible in relation to the decision-making process, as it allows to perform clustering based on various criteria, such as maximum mutual removal of clusters, proximity of geometric images of objects to the center of the cluster, other criteria. This is achieved by changing the calculation formula of the fitness function, which takes into account the necessary combination of clustering criteria without changing the structure of the algorithm. The algorithm is insensitive to initialization, as in the process of evolution of chromosomes through the use of genetic operators, the algorithm completely covers the whole set of acceptable solutions, which, in turn, provides high quality market segmentation.