A novel immune plasma-based automatic programming method for electricity price forecasting in the Turkish market
Engineering Applications of Artificial Intelligence, cilt.181, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 181
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.engappai.2026.115789
- Dergi Adı: Engineering Applications of Artificial Intelligence
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Compendex, INSPEC, Academic Search Ultimate (EBSCO), Engineering Source (EBSCO)
- Anahtar Kelimeler: Artificial Bee Colony Programming, Electricity price forecasting, Genetic programming, Immune plasma programming, Symbolic regression
- Sivas Cumhuriyet Üniversitesi Adresli: Evet
Özet
Electricity markets play an important role in modern electricity systems and promote competition through market mechanisms. Given this competition and the increasing demand for energy, it is important to ensure the reliability and overall efficiency of the markets. To this end, electricity price forecasting (EPF) helps market participants to optimize their strategies by minimizing risk. Forecasting models enable energy companies and other market players to increase their profitability and ensure the security of energy supply. It aims to protect consumers from price fluctuations and enables them to consume electricity at more stable prices. Many artificial intelligence subfields are used for EPF, including machine learning, deep learning and automatic programming (AP). AP-based models help to accurately describe the input–output relationships between systems using mathematical models and ensure the explainability of the model. In this paper, an improved version of immune plasma programming (IPP), called multi plasma immune programming (MPIP), is proposed. In the experiments, recursive feature elimination was first applied for feature selection, and the refined dataset was used to evaluate all methods. The proposed MPIP was compared with both conventional AP methods and deep learning models. According to the results, MPIP achieved the strongest overall performance across most of the reported evaluation settings, validating its effectiveness as a promising alternative with symbolic regression for EPF.