Evaluating the Predictive Performance of Monthly Inflation Rates in Sri Lanka using the Hybrid Model (HB)

Abstract

Aims/ objectives: This study develops and evaluates a novel hybrid model (HB) for forecasting monthly inflation rates in Sri Lanka, a country with a unique economic context, from 1988 to 2021. By integrating the Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Networks (ANNs), the study aims to overcome the limitations of traditional linear models in capturing the nonlinear patterns often observed in Sri Lankan economic data. Objectives: The study aims to assess the predictive accuracy of the HB model against established models, emphasizing its adaptability and robustness over a historically significant period. Methodology: Utilizing historical data, the study compares the HB model's forecasting performance with other established models, focusing on the Mean Absolute Percentage Error (MAPE) as a key metric of predictive accuracy. Results: The HB model demonstrates superior forecasting accuracy, with a notable reduction in MAPE to 7.10%, indicating its effectiveness in capturing the complexities of the Sri Lankan inflation trend. Conclusion: This study contributes to the field of economic forecasting by presenting a model that not only provides more accurate predictions but also adapts to the specific economic conditions of Sri Lanka. The findings have significant implications for economic planning and policy-making, highlighting the utility of hybrid forecasting models in developing economies.