Creation of an intelligent forecasting system to improve the efficiency of managing the liquidity and loan portfolio of a commercial bank
TL;DR
The article reveals the developed model's high accuracy of forecasting, and proposes recommendations on the integration of the system into the bank's business processes, responding to modern challenges of the financial sector.
Abstract
Subject. This article discusses liquidity management and forecasting of consumer credit sales in a commercial bank, and the need to switch from traditional forecasting methods to intelligent systems. Objectives. The study aims to develop and test a neural network forecasting model designed for short-term forecasting of consumer loan sales. Methods. For the study, we used methods of machine learning, Feature engineering techniques and the adaptive Adam algorithm. Results. The article reveals the developed model's high accuracy of forecasting, and proposes recommendations on the integration of the system into the bank's business processes. Conclusions. The use of neural network models to predict bank sales makes it possible to increase the bank's economic stability and minimize operational risks, responding to modern challenges of the financial sector.