The COVID-19 pandemic deepened understanding of e-commerce as an extremely promising sphere. Nowadays, even small businesses are widely using eshops and e-markets. Thus, small and medium-sized e-commerce companies need powerful, flexible recommender systems, which do not require significant

Application of predictive models on the basis of data mining confirmed its expediency in solving many economic problems. One of the crucial issues is the assessment of the borrower's creditworthiness on the basis of credit scoring models. This paper proposed an ensemble-based technique combining selected base classification models with business-specific feature selection add-on to increase the classification accuracy of real-life case of credit scoring. As the model limitations have been used easy-understandable algorithms on open-source software (R programming). The statistical results proved that hybrid approach for user-defined variables can be more than useful for ensemble binary classification model. It is shown that a great improvement can be reached by applying hybrid approach to feature selection process on additional variables (more descriptive ones that were built on initial features) for this real-life case with limited computational resources.