USE OF MACHINE LEARNING METHODS AND MICROSERVICE ARCHITECTURE IN PRODUCT ANALYTICS

Abstract

The purpose of this article is to present a customer classification methodology using a microservice architecture, which can be used to make marketing decisions and improve product management strategy and customer interaction. Customer classification is important to product analytics for a number of reasons, as it allows product manufacturers and businesses to understand their audience, adjust engagement strategies, and maximize the effectiveness of products and services. Methodology of research. The research methodology involves the use of the following methods: logical generalization, cluster analysis, analytical method (statistical analysis). Customer classification methodology is a systematic approach to grouping customers according to various characteristics or criteria to further provide personalized service and develop interaction strategies. The main purpose of such a classification is to learn about the variety of customers, their needs and characteristics in order to optimize communication and provide more effective services. Findings. The developed client classification microservice can be located both on a local host and on the Internet, certain parameters of the client are transferred to the model and it forms a forecast based on these data, which is essentially the number of the probable class to which this client should be assigned. Based on the analysis of the customer base, the construction of a clustering model, the Python programming language, and a set of specialized libraries, a microservice was built that will help classify new customers according to certain parameters. The use of microservices architecture facilitates the development, testing and maintenance of complex applications, providing flexibility and speed of response to changes in requirements and market conditions. However, proper planning, implementation, and management of microservices is key to avoiding the complexities associated with the growing number of services and their interactions. Practical value. The results of this study have practical implications for marketers and product management managers looking for tools to predict customer behavior. The proposed methodology and microservice can be used to create a decision support system in marketing activities.