A Dual-Bank Hybrid Predictive Model (DBHPM) for Financial
Forecasting
Samandarboy Sulaymanov1,∗
1Tashkent State University of Economics, Uzbekistan
Email: s.sulaymanov@tsue.uz
Abstract
Forecasting of the financial performance is significant mainly for the purpose of strategy formulation and
identification of potential problems in banking institutions. This paper presents a new model of a predictive
model for financial forecasting called the Dual-Bank Hybrid Predictive Model which consists of a Multiple
Linear Regression and Random Forest Regression. This model is also validated on two actual financial datasets
of Agrobank and NBU Bank from the year 2021 to 2025. It also relies on the analysis of such financial ratios
as Net profit, Equity, and Solvency which have been forecasted up to the year 2027. Specifically, while the
DBHPM consists of linear modeling through MLR in the first step, and then, nonlinear residuals thru RFR
in the second step of the analysis, the former provides increased generalizations and predictive strength as
compared to the later stage solely. The experimental results show that DBHPM minimizes MAE and RMSE
achieving the coefficient of determination (R2) amounting to 0.95 and above if compared to the models trained
independently. Statistical modelling shows that the two banks go up with Agrobank at approximately 1.18
billion sum and NBU Bank at 3.66 billion sum of the net profit by the end of 2027. The outlined hybrid model
presents the possibility of better predictive analytics financial modelling in the banking industry for purposes
of, decision-making, risk alertness, and economic forecast.
Keywords: Banking Sector Prediction; Agrobank; NBU Bank; Profitability Prediction; Machine Learning in
Finance; Strategic Financial Planning