Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Journal Sensi: Strategic of Education in Information System
Tia Wulandari, Anur Rahmah
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Performance Evaluation of ARIMA and LSTM Models with Product Inventory Demand in Production Companies Martono, Aris; Padeli, Padeli; Suhaepi, Muhamad Iip; Tia Wulandari, Anur Rahmah
Journal Sensi: Strategic of Education in Information System Vol 11 No 2 (2025): Journal SENSI
Publisher : UNIVERSITAS RAHARJA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33050/sensi.v11i2.4068

Abstract

This study aims to evaluate and compare the performance of two time series forecasting approaches: the classical statistical ARIMA model and the deep learning-based LSTM model, in the context of forecasting product inventory demand in a production company. The data used consists of historical daily demand records, totaling 100 and 200 records, which were analyzed to identify linear and non linear patterns. The ARIMA model was selected for its reliability in modeling stationary and seasonal data, while the LSTM model was utilized to capture complex temporal patterns through its layered neural network architecture. The test results using the MSE and RMSE metrics show that in both datasets, the ARIMA model has better prediction performance (100 records, RMSE=45.61% and 200 records, RMSE=44.72%) compared to LSTM, namely 100 records, RMSE=45.93% and 200 records, RMSE=49.54%. Although LSTM excels in handling non-linear dynamics, ARIMA outperformed it on data with linear. This study highlights the importance of selecting forecasting models based on data characteristics and suggests opportunities for future exploration of hybrid models. The theoretical and empirical foundations of this research are supported by the works of Hyndman & Athanasopoulos (2018), Hochreiter & Schmidhuber (1997), and Makridakis et al. (2018), which provide critical insight into predictive modeling for time series analysis.
Co-Authors Martono, Aris Padeli Padeli Suhaepi, Muhamad Iip