Performance Evaluation of ARIMA and LSTM Models to Handle Multi-Interventions in Automobile Production Forecasting

Firda Aulia Maghfiroh; Indahwati Indahwati; Asep Saefuddin

doi:10.55681/jige.v6i4.4694

Authors

Firda Aulia Maghfiroh Study Program on Statistics and Data Science, IPB University, Indonesia
Indahwati Study Program on Statistics and Data Science, IPB University, Indonesia
Asep Saefuddin Study Program on Statistics and Data Science, IPB University, Indonesia

DOI:

https://doi.org/10.55681/jige.v6i4.4694

Keywords:

Multi-Input Intervention, LSTM, ARIMA, Car Production

Abstract

Intervention refers to disturbances caused by internal or external variables, such as market changes, international events, or policy shifts. The dataset used in this study contains three intervention events, referred to as a multi-input intervention. The data consist of car production figures from PT Astra Daihatsu Motor obtained from the official GAIKINDO website. The forecasting task focuses on predicting PT Astra Daihatsu Motor’s production, which was influenced by three major interventions: policy changes in 2013, the impact of the COVID-19 pandemic in 2020, and the increase in SUV production in 2022. This study compares ARIMA and LSTM models for car production forecasting. The dataset covers monthly production data from January 2010 to June 2024, with a total of 174 observations. RMSE, MAPE, and SMAPE are employed as accuracy measures. Based on the testing data (May 2023–June 2024), the results show that the LSTM model outperforms ARIMA in capturing trend patterns, with lower error values of RMSE (4587.65), MAPE (10.37), and SMAPE (10.39), compared to ARIMA with RMSE (5059.48), MAPE (11.59), and SMAPE (10.50). Accordingly, LSTM represents a relevant and robust modeling alternative for production forecasting in operational decision-making, owing to its flexibility and strong capability in capturing complex data patterns.

Downloads

Download data is not yet available.

References

Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19(6), 716–723.

Aziza VN, Hilali FM, Maghfiroh FA, Notodiputro KA, & Angraini Y. (2023). Comparison of SARIMA Intervention and PROPHET Models for Forecasting The Number of Airline Passenger At Soekarno-Hatta International Airport. BAREKENG: J. Math. & App, Volume 17(4), 2107–2120. https://doi.org/10.30598/barekengvol17iss4pp2107-2120

Bandara, K., Bergmeir, C., & Smyl, S. (2020). Forecasting across time series databases using recurrent neural networks. International Journal of Forecasting, 36(3), 855–873.

Benidis, K., Rangapuram, S. S., Flunkert, V., et al. (2020). Neural forecasting: Introduction and literature overview. IEEE Signal Processing Magazine, 37(6), 97–110.

Bergmeir, C., Hyndman, R. J., & Benítez, J. M. (2018). Bagging exponential smoothing methods. International Journal of Forecasting, 32(2), 303–312.

Box, G. E. P., & Jenkins, G. M. (1970). Time Series Analysis: Forecasting and Control. Holden-Day.

Cryer, J. D., & Chan, K. S. (2008). Time Series Analysis: With Applications in R (2nd ed.). Springer.

Dubey A K, Kumar A, Díaz V G, Sharma A K, & Kanhaiya K. (2021). Study And Analysis of SARIMA And LSTM in Forecasting Time Series Data. Sustainable Energy Technologies and Assessments, Vol 47. https://doi.org/10.1016/j.seta.2021.101474

Enders, W. (2015). Applied Econometric Time Series (4th ed.). Wiley.

Fildes, R., Ma, S., & Kolassa, S. (2019). Retail forecasting: Research and practice. International Journal of Forecasting, 35(1), 1–9.

GAIKINDO. (n.d.). Indonesian Automobile Industry Data. Gabungan Industri Kendaraan Bermotor Indonesia. Retrieved February 2, 2024, from https://www.gaikindo.or.id/indonesian-automobile-industry-data/

Hamilton, J. D. (2018). Why you should never use the Hodrick–Prescott filter. Review of Economics and Statistics, 100(5), 831–843.

Iswari A, Angraini Y, & Masjkur M. (2022). Comparison of The SARIMA Model and Intervention in Forecasting The Number of Domestic Passengers at Soekarno-Hatta International Airport. Indonesian Journal of Statistics and Its Applications, Vol 6 No 1(1), 132–146. https://doi.org/10.29244/ijsa.v6i1p132-146

Kilian, L., & Lütkepohl, H. (2017). Structural Vector Autoregressive Analysis. Cambridge University Press.

Lorensya, Ambarwati L, & Mahatma Y. (2022). Analisis Model Intervensi Multi Input Fungsi Pulse dan Fungsi Step dengan Deteksi Outlier terhadap Harga Saham. Jurnal EurekaMatika, Volume 10(Issue 1), 61–72.

Masrawanti, Wahyuningsih S, & Hayati M N. (2019). Analisis Model Intervensi Fungsi Step Ganda untuk Peramalan Inflasi Indonesia (Studi Kasus: Inflasi Indonesia Tahun Analysis of Intervention Models Double Step Function for Forecasting Indonesia’s Inflation (Case Study: Indonesia’s Inflation in 2009-2017). Jurnal EKSPONENSIAL, 10 Nomor 2(2), 119–126.

Papacharalampous, G., & Tyralis, H. (2018). Evaluation of machine learning algorithms for time series forecasting. Water Resources Research, 54(8), 5597–5613.

Primawati A, Sitanggang I S, Annisa, & Astuti D A. (2023). Perbandingan Kinerja Lstm dan Prophet untuk Prediksi Deret Waktu (Studi Kasus Produksi Susu Sapi Harian). JEPIN (Jurnal Edukasi Dan Penelitian Informatika), Vol. 9, No. 3, 428–435.

Ribeiro, M. H. D. M., Coelho, L. S., & dos Santos, A. G. (2020). Multi-step ahead forecasting with deep learning. Neurocomputing, 396, 173–184.

Sautomo S, & Pardede HF. (2021). Prediksi Belanja Pemerintah Indonesia Menggunakan Long Short-Term Memory (LSTM). Jurnal RESTI (Rekayasa Sistem Dan Teknologi Informasi), 5(1), 99–106. https://doi.org/10.29207/resti.v5i1.2815

Schaffer AL, Dobbins TA, & Pearson SA. (2021). Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models: a guide for evaluating large-scale health interventions. BMC Medical Research Methodology, 21(1). https://doi.org/10.1186/s12874-021-01235-8

Shumway, R. H., & Stoffer, D. S. (2017). Time Series Analysis and Its Applications (4th ed.). Springer.

Sirisha U M, Belavagi M C, & Attigeri G. (2016). Profit Prediction Using ARIMA, SARIMA and LSTM Models in Time Series Forecasting: A Comparison. IEE Access, Vol 4, 1–13. https://doi.org/10.1109/ACCESS.2017.DOI

Sohibien GPD. (2018). Analysis of The Effect of Fuel Price Policy on Jakarta Inflation by Using Multi-Input Intervention Model. AIP Conference Proceedings, 1–11. https://doi.org/10.1063/1.5054529

Taylor, J. W., & Letham, B. (2018). Forecasting at scale. The American Statistician, 72(1), 37–45.

Wei W W S. (2006). Time Series Analysis Univariate and Multivariate Methods (Second).

Wiladibrata M I, & Rifai N A K. (2022). Peramalan Produksi Mobil Menggunakan Metode Double Exponential Smoothing Dengan Algoritma Golden Section. Bandung Conference Series: Statistics, Vol 2 No 2, 507–511. https://doi.org/10.29313/bcss.v2i2.4776

Zhang, Y., & Yang, Q. (2021). A survey on multi-task learning. IEEE Transactions on Knowledge and Data Engineering, 34(12), 5586–5609.