Skip to main navigation Skip to search Skip to main content

Prediction of Surabaya City's inflation rate based on staple food price forecast using Multivariate Long Short-Term Memory (LSTM)

  • Department of Information Systems
  • Institut Teknologi Sepuluh Nopember

Research output: Contribution to journalConference articlepeer-review

Abstract

Inflation is a crucial economic indicator affecting purchasing power, influenced by staple food price fluctuations. Previous studies highlight rice, red chilies, and shallots as key contributors to short-term inflation. Surabaya, a city with the highest inflation in East Java in October 2023, requires effective control measures. This study analyzes historical weekly data (August 2017 - December 2023) of the staple food (rice, shallot, and red chili) price and proposed a multivariate LSTM model for predicting the weekly inflation rate of Surabaya based on the staple food price forecast. The proposed LSTM model can forecast the staple food price better than ARIMA model, with a difference in price error ranging from IDR 800 to around IDR 20,000. The staple food price forecast results then become the input of the proposed multivariate LSTM model. The experimental results show that the proposed multivariate LSTM model predicts weekly inflation rates with an RMSE of 0.1854 and a MAPE of 48.31%, outperforming univariate LSTM and ARIMA model.

Original languageEnglish
Pages (from-to)1619-1628
Number of pages10
JournalProcedia Computer Science
Volume284
DOIs
Publication statusPublished - 2026
Event8th Information Systems International Conference, ISICO 2025 - Hybrid, Bali, Indonesia
Duration: 4 Aug 20256 Aug 2025

Keywords

  • Inflation
  • LSTM
  • Machine Learning
  • Multivariate Forecasting
  • Staple Food Price

Fingerprint

Dive into the research topics of 'Prediction of Surabaya City's inflation rate based on staple food price forecast using Multivariate Long Short-Term Memory (LSTM)'. Together they form a unique fingerprint.

Cite this