EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

Comparative Evaluation of Random Forest and Support Vector Machine for Interpretable Breast Cancer Prediction

Herwis Gultom, Indra Kristianto

bit-Tech · 10.04.2026 · 8 (3) · 4132-4143

DOI: 10.32877/bt.v8i3.3821 ↗

Kurz gefasst

Fachartikel in der Zeitschrift bit-Tech, online erschienen am 10.04.2026. Verfasst von Herwis Gultom und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Breast cancer prediction requires classification models that are not only accurate but also clinically meaningful in minimizing missed malignant cases. This study addresses the research question of whether Random Forest (RF) and Support Vector Machine (SVM) differ meaningfully in sensitivity-oriented breast cancer classification when evaluated under a consistent empirical benchmarking framework. Using the Breast Cancer Wisconsin Diagnostic Dataset from Kaggle, comprising 569 instances and 30 numerical diagnostic features, the study implemented a supervised machine learning workflow involving data cleaning, label encoding, StandardScaler-based feature standardization, stratified 80:20 train–test partitioning, model training, and hyperparameter optimization. Performance was assessed using accuracy, precision, recall, F1-score, confusion matrix analysis, and Area Under the Curve (AUC). The SVM model achieved 97.36% accuracy, 100% precision, 92.85% recall, 96.29% F1-score, and 99.54% AUC, whereas RF achieved 96.49% accuracy, 100% precision, 90.47% recall, 95.00% F1-score, and 99.60% AUC. The primary contribution is therefore positioned as empirical benchmarking rather than a new explainable AI framework. SVM produced fewer false negatives, indicating stronger sensitivity for malignant-case detection at the selected decision threshold, while RF provided complementary feature-importance evidence for identifying influential diagnostic variables. These findings clarify the trade-off between sensitivity-driven predictive reliability and model-specific interpretability, suggesting that SVM is preferable for reducing missed malignant cases, whereas RF remains useful when transparent feature-level insight is required.

Zitieren

APA

Gultom, H., & Kristianto, I. (2026). Comparative Evaluation of Random Forest and Support Vector Machine for Interpretable Breast Cancer Prediction. bit-Tech, 8(3), 4132-4143. https://doi.org/10.32877/bt.v8i3.3821

Vancouver

Gultom H, Kristianto I. Comparative Evaluation of Random Forest and Support Vector Machine for Interpretable Breast Cancer Prediction. bit-Tech. 2026;8(3):4132-4143. doi:10.32877/bt.v8i3.3821

BibTeX
@article{Gultom2026,
  author = {Gultom, Herwis and Kristianto, Indra},
  title = {Comparative Evaluation of Random Forest and Support Vector Machine for Interpretable Breast Cancer Prediction},
  journal = {bit-Tech},
  year = {2026},
  volume = {8},
  number = {3},
  pages = {4132--4143},
  doi = {10.32877/bt.v8i3.3821}
}
Kostenlos starten →