Development of Predictive Analytical Models for Profiling Internet-users Using Machine Learning Methods

The present Master’s thesis is devoted to the development of predictive analytics models to profile users in online shops using machine learning methods. The study considers the key activities of an e-commerce enterprise and the process of typical user profiling, as well as the possibilities of machine learning application to in the development of predictive models. The predictive model was developed on the basis of naive Bayes method and tested on a test dataset of an online shop using Python 3.7. To achieve the goal, the study of current technological trends in the application of machine learning methods in marketing was conducted. Global trends in the development of e-commerce, as well as the specifics of the domestic market were analyzed. A review of the methods and tools of business analytics in profiling tasks is carried out. An analysis of methods and profiling models is performed. The requirements for the developed profiling model are identified and the possible study limitations of study are evaluated. Expert assessment of business metrics in the field of marketing based on existing methodologies, as well as quality assessment of the developed model, are carried out. The study is of interest both for researchers in the field of digital marketing and for specialists in the field of marketing and development.