Application of the ARIMA Model in Vertica for Wind Speed Forecasting

• Gabriel Ciriaco Fornitano Federal University of Itajubá
• Flávio Belizário Mota Discover Strategic Decision Solutions
• Vanessa Cristina Oliveira de Souza Federal University of Itajubá
• Arcilan Assireu Federal University of Itajubá
• Melise Maria Veiga de Paula Federal University of Itajubá

DOI: https://doi.org/10.5753/sbbd.2025.247487

Abstract

This study presents an exploratory analysis of applying the ARIMA model directly within the Vertica database for wind speed forecasting. A dataset from the EOSOLAR project was used, containing vertical wind profile measurements from the coastal region of Maranhão, Brazil. The evaluation considered the RMSE (Root Mean Squared Error) and MAE (Mean Absolute Error) metrics across 105 trained models. The study investigated whether the in-database ARIMA approach in Vertica could provide efficient modeling for wind speed forecasting. The results showed that models with low complexity achieved good predictive performance.

Keywords: In-Database Machine Learning, Séries Temporais, Previsão da Velocidade do Vento, ARIMA

References

Assireu, A. T., Fisch, G., Carvalho, V. S. O., Pimenta, F. M., de Freitas, R. M., Saavedra, O. R., Neto, F. L. A., Júnior, A. R. T., Oliveira, D. Q., Lopes, D. C. P., de Lima, S. L., Marcondes, L. G. P., and Rodrigues, W. K. S. (2024). Sea breeze-driven effects on wind down-ramps: Their implications for wind farms along the north-east coast of brazil. Energy, 294:130804.

Boettiger, C. (2015). An introduction to docker for reproducible research. ACM SIGOPS Operating Systems Review, 49(1):71–79.

Box, G. E. P., Jenkins, G. M., Reinsel, G. C., and Ljung, G. M. (2015). Time Series Analysis: Forecasting and Control. Wiley, 5 edition.

Chastre, C. and Lúcio, V. (2012). Torres pré-fabricadas de betão para suporte de turbinas eólicas. In Estruturas pré-moldadas no mundo – Aplicações e comportamento estrutural, pages 91–106. Universidade NOVA de Lisboa.

Cielen, D., Meysman, A. D. B., and Ali, M. (2021). Data Science: Principles and Practice. Manning Publications.

Elsaraiti, M. and Merabet, A. (2021). A comparative analysis of the arima and lstm predictive models and their effectiveness for predicting wind speed. Energies, 14(20):6782.

Epstein, B. and Roberts, P. (2022). Accelerate Machine Learning with a Unified Analytics Architecture. O’Reilly Media, Inc., Sebastopol, CA, USA.

Fard, A., Zhang, B., Katepalli, K., Stonebraker, M., and Rundensteiner, E. A. (2020). Vertica-ml: Distributed machine learning in vertica database. In Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data, pages 755–768. ACM.

Grigonytė, E. and Butkevičiūtė, E. (2016). Short-term wind speed forecasting using arima model. Energetika, 62(1–2):17–26.

Hyndman, R. J. and Athanasopoulos, G. (2021). Forecasting: Principles and Practice. OTexts, Melbourne, Australia, 3 edition. Accessed on March 26, 2025.

Lamb, A., Fuller, M., Varadarajan, R., Tran, N., Vandiver, B., Doshi, L., and Bear, C. (2012). The vertica analytic database: C-store 7 years later. Vertica Systems, An HP Company.

Liu, X., Lin, Z., and Feng, Z. (2021). Short-term offshore wind speed forecast by seasonal arima-a comparison against gru and lstm. Energy, 227:120492.

Lustosa, H., Costa, F., Guimarães, J., and de Oliveira, D. (2020). Savime: An array dbms for simulation analysis and ml models predictions. In International Conference on Database and Expert Systems Applications, pages 357–367. Springer.

Raschka, S., Patterson, J., and Nolet, C. (2020). Machine learning in python: Main developments and technology trends in data science, machine learning, and artificial intelligence. Information, 11(4):193.

Salman, A. G. and Kanigoro, B. (2021). Visibility forecasting using autoregressive integrated moving average (arima) models. Procedia Computer Science, 181:586–593.

Vertica (2025). Arima - vertica 25.1.x documentation. Accessed: March 26, 2025.