2026-05-10T08:05:46.349Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:9559 2024-12-03T20:10:47Z 3

Oil Price Prediction Approach Using Long Short-Term Memory Network Tuned by Improved Seagull Optimization Algorithm, Chapter in AIS Algorithms for Intelligent Systems: ICSISCET 2022: Artificial Intelligence and Sustainable Computing, Springer 2023 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/9559 https://link.springer.com/chapter/10.1007/978-981-99-1431-9_20 L. Jovanovic N. Bacanin A. Jovancai D. Jovanovic D. Singh M. Antonijevic M. Zivkovic I. Strumberger The prices of crude oil have a considerable effect on global economic activity. This makes forecasting crude oil prices one of the most interesting and potentially very lucrative research topics. However, as the price of crude oil is very volatile it also represents a very difficult and complex topic, spanning multiple fields of research. Artificial intelligence (AI) has previously shown great promise when tackling complex problems, and by leveraging long short-term memory (LSTM) networks this work attempts to address the challenges of forecasting crude oil prices. Moreover, the conducted research focuses on optimizing the forecasting process and improving result accuracy by selecting the best possible hyper-parameters through a process known as hyper-parameter tuning. This process is tackled by applying swarm intelligence. This work proposed enhancements of the original seagull optimization algorithm (SOA) dubbed the improved seagull optimization algorithm (ISOA). The improved algorithm is applied to selecting optimal LSTM hyper-parameters for crude oil price prediction. Testing conducted on real-world WTI price data suggests that the proposed LSTM-ISOA approach outperforms other state-of-the-art metaheuristics tackling the same optimization task as well as the original SOA. bookPart Springer, Singapore 253 265 10.1007/978-981-99-1431-9_20 AIS Algorithms for Intelligent Systems: ICSISCET 2022: Artificial Intelligence and Sustainable Computing