Time Series Prediction of Temperature Using Seasonal ARIMA and LSTM Models

Yıl 2023, Cilt: 9 Sayı: 3, 574 - 584, 01.01.2024

Hakan Koçak

Öz

Precise quantitative understanding and monitoring of temperature is indispensable due to its tremendous impact on almost every aspect of our lives. This work investigates prediction capabilities of two machine learning techniques, namely Autoregressive Integrated Moving Average (ARIMA) and Long Short Term Memory (LSTM) and compares them in predicting monthly mean temperature time series data for a weather station in Ankara, Türkiye from January 2010 to March 2023. The comparison of forecasting performance was based on mean square error (MSE), root mean square error (RMSE) and mean absolute error (MAE). The results showed that both models can capture the variations of time series data. Both models exhibited reasonably good performance in predicting monthly mean temperature, but the ARIMA model gave the least forecast error compared to the LSTM model.

Anahtar Kelimeler

Time series prediction, monthly mean temperature, ARIMA, SARIMA, LSTM, Autocorrelation Function (ACF)

ARIMA ve LSTM Modelleri ile Sıcaklık Zaman Serileri Tahmini

Öz

Hava sıcaklığının insan hayatının hemen her alanındaki büyük öneminden dolayı sıcaklığın nicel olarak anlaşılması ve takip edilmesi oldukça elzemdir. Bu çalışmada Ankara’da bulunan bir meteorolojik hava istasyonundan elde edilen Ocak 2010 ila Mart 2023 tarihleri arasındaki gözlem verileri kullanarak ARIMA (Autoregressive Integrated Moving Average) ve LSTM (Long Short Term Memory) makine öğrenmesi metotlarıyla aylık ortalama hava sıcaklığını kestirimi yapılmış ve bu metotların sıcaklık tahmini konusundaki başarısı karşılaştırılarak irdelenmiştir. Modellerin tahmin başarıları Ortalama Karesel Hata (OKH), Ortalama Karesel Hata (KOKH) ve Ortalama Mutlak Hata performans metrikleri kullanarak yapılmıştır. Araştırma sonucunda, aylık ortalama sıcaklık kestiriminde her iki modelin de iyi derecede performans gösterdiği görülmekle birlikte ARIMA modelinin LSTM modeline göre hata oranının daha az olduğu, dolayısıyla daha iyi performans gösterdiği ortaya çıkmıştır.

Anahtar Kelimeler

— Zaman serileri tahmini, aylık ortalama yağış, ARIMA, SARIMA, LSTM, Autocorrelation Function (ACF)

Kaynakça

• [1] M. Murat, I. Malinowska, M. Gos, J. Krzyszczak, “Forecasting daily meteorological time series using ARIMA and regression models,” International Agrophysics, vol. 32, no. 2, pp. 253–264, 2018. doi:10.1515/intag-2017-0007
• [2] S. Siami-Namini, N. Tavakoli and A. Siami Namin, "A Comparison of ARIMA and LSTM in Forecasting Time Series," in Proc. of the 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), 17-20 Dec. 2018, Orlando, FL, USA [Online]. Available: IEEE Xplore, https://ieeexplore.ieee.org. [Accessed: 7 Agu. 2023]
• [3] P. Liu, ‘Time Series Forecasting Based on ARIMA and LSTM’, in Proceedings of the 2022 2nd International Conference on Enterprise Management and Economic Development (ICEMED 2022), 2022, pp. 1203–1208.
• [4] A. Parasyris, G. Alexandrakis, G.V. Kozyrakis, K. Spanoudaki, N.A. Kampanis, “Predicting Meteorological Variables on Local Level with SARIMA, LSTM and Hybrid Techniques”, Atmosphere 2022, vol. 13, no. 878, 2022. doi:10.3390/atmos13060878
• [5] K. Tadesse, O. Megersa, and M. Dinka, “The SARIMA model-based monthly rainfall forecasting for the Turksvygbult Station at the Magoebaskloof Dam in South Africa”, Journal of Water and Land Development, vol. 53, pp. 100-107, 2022. doi:10.24425/jwld.2022.140785
• [6] C. Chatfield, The analysis of time series: An introduction, 6th ed. London, UK: Chapman & Hall/CRC, 2004.
• [7] I. S. Rahayu, E. C. Djamal, R. Ilyas, “Daily Temperature Prediction Using Recurrent Neural Networks and Long-Short Term Memory”, in Proceedings of the 5th NA International Conference on Industrial Engineering and Operations Management, August 10 – 14, 2020, Detroit. Michigan, USA [Online]. Available: ieomsociety.org, http://www.ieomsociety.org. [Accessed: 10 Apr. 2023].
• [8] J. Choi, D. C. Roberts, and E. Lee, “Forecasting oil production in North Dakota using the seasonal autoregressive integrated moving average (S-ARIMA)”, Natural Resources, vol. 6, no. 1, pp. 16–26, 2015.
• [9] A. Can and A. T. Atimtay, “Time series analysis of mean temperature data in Turkey,” Applied Time Series, vol. 4, no. 20, 2002.
• [10] S. Zakaria, N. Al-Ansari, S. Knutsson, and T. Al-Badrany, “ARIMA models for weekly rainfall in the semi-arid Sinjar district at Iraq,” Journal of Earth Science Geotechnical Engineering, vol. 2, no. 3, pp. 25-55, 2012.
• [11] P. Chen, A. Niu, D. Liu, W. Jiang, and B. Ma, “Time series forecasting of temperatures using SARIMA: An example from Nanjing,” IOP Conference Series: Materials Science and Engineering, vol. 394, 2018. doi:10.1088/1757- 899X/394/5/052024
• [12] D. K. Dwivedi, G. R. Sharma and S. S. Wandre, “Forecasting mean temperature using SARIMA Model for Junagadh City of Gujarat,” IJASR, vol. 7, no. 4, pp. 183–194, Jan. 2017. doi:10.24247/ijasraug201723
• [13] J. Asha, S. S. Kumar, and S. Rishidas, ‘Forecasting performance comparison of daily maximum temperature using ARMA based methods’, Journal of Physics: Conference Series, vol. 1921, no. 1, p. 012041, May 2021. doi:10.1088/1742-6596/1921/1/012041
• [14] K. M. S. A. Hennayake, R. Dinalankara, and D. Y. Mudunkotuwa, “Machine Learning Based Weather Prediction Model for Short Term Weather Prediction in Sri Lanka,” International Journal of Multidisciplinary Studies (IJMS), vol 9, no. I, 2022. doi:10.4038/ijms.v9i1.159
• [15] K. N. Mitu, K. Hasan, “Modelling and Forecasting Daily Temperature Time Series in the Memphis, Tennessee,” International Journal of Environmental Monitoring and Analysis, vol. 9, no. 6, pp. 214-221, Dec. 2021. doi:10.11648/j.ijema.20210906.17
• [16] T. Dimri, S. Ahmad, and M. Sharif, “Time series analysis of climate variables using seasonal ARIMA approach,” Journal of Earth System Science volume, vol. 129, 2020. doi:10.1007/s12040-020-01408-x
• [17] A. Gangshetty, G. Kaur, U. S. Malunje, “Time Series Prediction of Temperature in Pune using Seasonal ARIMA Model,” International Journal of Engineering Research & Technology (IJERT), vol. 10, Issue 11, November-2021.
• [18] D. T. Hoang, Pr. L. Yang, L. D. P. Cuong, P. D. Trung, N. H. Tu, L. V. Truong, T. T. Hien, V. T. Nha, “Weather prediction based on LSTM model implemented AWS Machine Learning Platform,” International Journal for Research in Applied Science & Engineering Technology (IJRASET), vol 8, no. V, May 2020. doi:10.22214/ijraset.2020.5046
• [19] "What Is ARIMA Modeling?," Master's in Data Science, edX.org, 2023. [Online]. Available: https://www.mastersindatascience.org/learning/statistics-data-science/what-is-arima-modeling. [Accessed: Apr. 10, 2023]
• [20] R. J. Hyndman and G. Athanasopoulos, “Forecasting: principles and practice”, 2013.
• [21] R. Dolphin. “LSTM Networks | A Detailed Explanation | A Comprehensive Introduction to LSTMs”. Towards Data Science, medium.com, Oct 21, 2020. [Online]. Available: https://towardsdatascience.com/lstm-networks-a-detailed-explanation-8fae6aefc7f9. [Accessed: May 11, 2023].
• [22] A. I. Arasu, M. Modani, and N. R. Vadlamani, “Application of Machine Learning Techniques in Temperature Forecast”, in Proc. of the 21st IEEE International Conference on Machine Learning and Applications (ICMLA), 2022, Nassau, Bahamas, pp. 513-518, doi:10.1109/ICMLA55696.2022.00083