Work

Time Series Forecasting with Ensemble Methods for Inflation in Indonesia

Year: 2026

Type: article

Abstract: Inflation is an economic problem experienced by almost all countries in the world so inflation needs to be controlled. Controlling inflation requires accurate inflation forecasting. In this study, the ensemble method was used for the modeling. Three ensemble methods were tested, namely simple average, trimmed average and variance-based. The individual models applied in this research are based on Machine Learning, which include Long Short-Term Memory (LSTM), Extreme Learning Machine (ELM), and Feed Forward Neural Network (FFNN). In selecting inputs for individual models, statistical tools were used, namely correlation analysis and partial autocorrelation function. Among the three individual models, the best model was the FFNN model. While the best ensemble method was variance-based. The ensemble method can reduce the RMSE of the best individual model.

Source: TEM Journal

DOI: 10.18421/tem153-33

Authors: Eni Sumarminingsih, Rahma Fitriani, Aqsa Yudhistira Redi, Satriadi Putra Santika, Muhammad Fahmi Fauzan, Natasha Aulia

Institutions: University of Brawijaya, Binus University

Language: English

Topic: Forecasting Techniques and Applications

Subfield: Management Science and Operations Research

Field: Decision Sciences

Domain: Social Sciences

Open Access status: gold

Estimated APC paid: $647

Funder: Universitas Brawijaya