COMPARISON OF HOLT-WINTER EXPONENTIAL SMOOTHING AND LSTM METHODS ON RAINFALL FORECASTING IN SUBAK ABIAN WANASARI KENJUNG

Ni Made Wahyu Citrayanti; I Komang Arya Ganda Wiguna; Aniek Suryanti Kusuma; I Gede Iwan Sudipa; Ni Made Mila Rosa Desmayani

Authors

Ni Made Wahyu Citrayanti Institut Bisnis dan Teknologi Indonesia
I Komang Arya Ganda Wiguna Institut Bisnis dan Teknologi Indonesia
Aniek Suryanti Kusuma Institut Bisnis dan Teknologi Indonesia
I Gede Iwan Sudipa Institut Bisnis dan Teknologi Indonesia
Ni Made Mila Rosa Desmayani Institut Bisnis dan Teknologi Indonesia

Keywords:

holt-winter, LSTM, forecasting, method comparison

Abstract

Rainfall is important in the agricultural sector, as rainfall uncertainty affects planting schedules. Therefore, forecasting is necessary to help farmers to plan agricultural activities more efficiently, and reduce the risk of losses caused by weather fluctuations. Some methods can be used, two of which are Holt-Winters and LSTM. To test methods, a case study of one of the subaks in the Catur Village, namely Subak Abian Wanasari Kenjung, was taken. The data used is monthly rainfall data of Catur Village, obtained from BMKG Region III Denpasar. In this research, the Holt-Winter Multiplicative produces a MAPE of 47.88%, the Holt Winter Additive produces a MAPE of 47.69% and the LSTM produces a MAPE of 79.78%. From the test results, Holt-Winter Additive is the method that produces the lowest MAPE, namely 47.69%, so the Holt-Winter Additive is the most accurate and most suitable method for use in predicting future rainfall. The results of this study can help farmers optimize planting schedules and reduce weather-related risks. Stakeholders such as governments and agricultural organizations can also utilize rainfall predictions for efficient planning of water, fertilizer, and seed distribution, supporting timely decision-making. Thus, rainfall forecasts can enhance the accuracy and timeliness of decision-making processes.

References

Agustina, D. (2024). ADDITIVE HOLT-WINTERS METHOD FOR FORECASTING GROSS REGIONAL DOMESTIC PRODUCT AT CONSTANT. 18(4), 2737–2746.

Amarachi, O. G., Ayuba, H., & Ovye, A. D. (2024). Effects of rainfall and temperature on rice production in nigeria. Journal of Built Environment and Geological Research, 04(4), 65–80.

Arora, P., & Balyan, A. (2023). Comparative Analysis of LSTM, Encoder-Decoder and GRU Models for Stock Price Prediction. Lecture Notes in Electrical Engineering, 968, 399–410. https://doi.org/10.1007/978-981-19-7346-8_34

Evania, C. Della, & Siregar, B. (2024). Comparison of Holt Winter ’ s and SARIMA Methods on the data of the Number of Foreign Tourist Visits in Bali Province. 7(2), 350–359.

Heatubun, A. B., & Legowo, P. S. (2024). Anticipate risk and uncertainty through price forecasting strategies : Case study of curly red chili prices in traditional markets in several cities in West Java Province. 147–155.

Huang, Z., & Yi, Y. (2024). Short-Term Load Forecasting for Regional Smart Energy Systems Based on Two-Stage Feature Extraction and Hybrid Inverted Transformer. Sustainability, 16(17), 7613. https://doi.org/10.3390/su16177613

Kamisan, N. A. B., Norrulashikin, S. M., & Joe, T. T. (2024). Application of time series models to rainfall forecasting in Senai, Johor. AIP Conference Proceedings, 2895(1). https://doi.org/10.1063/5.0193354

muhamad, azhee, & Mohammed, aree. (2022). Improved Rnn Model for Real-Time Human Activity Recognition. SSRN Electronic Journal, 65(9), 5220–5238. https://doi.org/10.2139/ssrn.4105116

Olcay, K., Giray Tunca, S., & Arif Ozgur, M. (2024). Forecasting and Performance Analysis of Energy Production in Solar Power Plants Using Long Short-Term Memory (LSTM) and Random Forest Models. IEEE Access, 12(August), 103299–103312. https://doi.org/10.110A9/ACCESS.2024.3432574

Samson, T. K., & Aweda, F. O. (2024). Forecasting Rainfall in Selected Cities of Southwest Nigeria: A Comparative Study of Random Forest and Long Short-Term Memory Models. Acadlore Transactions on Geosciences, 3(2), 79–88. https://doi.org/10.56578/atg030202

Sari, I. K., Liyantono, & Solahudin, M. (2024). Comparative study of oil palm productivity estimation models: Multilinear regression and holt-winters approach. IOP Conference Series: Earth and Environmental Science, 1386(1). https://doi.org/10.1088/1755-1315/1386/1/012031

Syarifah, A., Mahdiyyah, A. A., & Prastawa, H. (2023). Usulan Perencanaan Peramalan (Forecasting) & Safety Stock Bahan Bakar Diesel Fuel Pada Unit Kerja Pt Badak Ngl Menggunakan Metode Time Series (Studi Kasus: Maintenance Departement Pt Badak Ngl Bontang). 1–8.

Triyana Muliawati. (2024). Analysis of Rainfall Prediction in Lampung Province using the Exponential Smoothing Method. International Journal of Scientific Research in Science, Engineering and Technology, February, 232–240. https://doi.org/10.32628/ijsrset2411127

Wang, Y., Huang, Y., Chen, D., Wang, L., Ye, L., & Shen, F. (2024). Prox-STA-LSTM: A Sparse Representation for the Attention-Based LSTM Networks for Industrial Soft Sensor Development. IEEE Access, 12(April), 80633–80645. https://doi.org/10.1109/ACCESS.2024.3409899

Zoremsanga, C., & Hussain, J. (2024). An Evaluation of Bidirectional Long Short-Term Memory Model for Estimating Monthly Rainfall in India. Indian Journal Of Science And Technology, 17(18), 1828–1837. https://doi.org/10.17485/ijst/v17i18.2505