Time series based financial forecasting models
- Autores: Ali Kipel, Cüneyt Yurt, Yıldırım Adigüzel, Zehra Sude Sari, M. Fatih Akay
- Localización: International Journal of Professional Business Review: Int. J. Prof.Bus. Rev., ISSN 2525-3654, ISSN-e 2525-3654, Vol. 10, Nº. 3, 2025 (Ejemplar dedicado a: Continuous publication; e05371)
- Idioma: inglés
- Enlaces
-
Resumen
Objective: The purpose of this study is to develop financial component forecasting models that enable more accurate financial planning, allowing businesses to gain a competitive advantage.
Theoretical Framework: Time series-based Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM), Multi-Quantile Recurrent Neural Network (MQRNN), and Autoregressive Integrated Moving Average (ARIMA) models have been developed for forecasting financial components.
Method: Forecasting models have been developed for two different months (June and July) using a dataset containing 291 rows of weekly data from 01.01.2017 to 31.07.2022. The dataset includes data for five different finanacial components including sales, purchase, cash payment, cash collection, and card collection. The performance of the models has been evaluated using Mean Absolute Percentage Error (MAPE).
Results and Discussion: The MAPE values obtained with the developed forecasting models range from 2.44% to 26.57%. The CNN-LSTM model exhibits the highest predictive performance among the evaluated models.
Research Implications: This research makes significant contributions to financial forecasting and planning by highlighting effectiveness of time series methods. It demonstrates that models developed using CNN-LSTM, MQRNN, and ARIMA perform differently for various financial components, with each model excelling under specific conditions. Practically, these models help businesses improve financial planning, optimize costs, and enhance profit margins.
Originality/Value: This study contributes to literature by evaluating the effectiveness of time series methods in financial forecasting models. The findings are applicable in areas such as financial services, retail, and business strategy, offering value for financial risk management, sales forecasting, and long-term decision-making.
