Forecasting Business Metrics Using Machine Learning Methods Considering Market Conditions and Business Constraints in the Corporate Banking Sector

This research focuses on product management in the corporate division of a bank using data-driven approaches. It examines the processes of forecasting financial metrics for budgeting in the corporate division, as well as methods for addressing these tasks using machine learning, considering market conditions and constraints. The study aims to develop a reliable forecasting tool tailored to the needs of corporate banking. Integrating data on market conditions and internal constraints enhances the practical utility of the research and the tool developed. The final practical result of the research is a web application for managers in the selected bank domain and a Python library for analysts and data scientists, enabling work with hierarchical time series. Thus, the research provides a platform for creating easily interpretable forecasts, facilitating more effective decision-making and strategic planning in a dynamic business environment. Keywords: Business metric forecasting, business modeling, advanced data analytics, time series prediction, financial metric decomposition.