Machine Learning Classification of A Fuel Cell: A Comparative Study Using B0005, B0006, B0018, and B0007 Derived Data

Akkaya S., Hayber Ş. E., Uyar M.

International Conference on Energy Systems (ICES 2025), İstanbul, Türkiye, 11 - 14 Mayıs 2025, ss.297-300, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.297-300
  • Bursa Uludağ Üniversitesi Adresli: Evet

Özet

The efficiency of various machine learning methods on fuel cells is extremely critical for the classification of fuel cells with high accuracy. Due to the similar characteristics between fuel cells and conventional batteries, such as Li-Ion, a large FC dataset was created using NASA’s B0005, B0006, B0018, and B0007 battery datasets. Five classification algorithms (Decision Tree, Naive Bayes, kNN, LDA, and SVM) were compared regarding performance on the FC dataset using four metrics (Accuracy, Precision, Recall, and F1 Score). The obtained values ​​show that the Decision Tree method provides quite satisfactory results in all metrics and is an effective ML method in FC classification.