EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

The Comparison of Extreme Machine Learning and Hidden Markov Model Algorithm in Predicting The Recurrence Of Differentiated Thyroid Cancer Using SMOTE

Nor Aida, Triando Hamonangan Saragih, Dwi Kartini, Radityo Adi Nugroho, Dodon Turianto Nugrahadi

Journal of Electronics, Electromedical Engineering, and Medical Informatics · 16.09.2024 · 6 (4) · 429-444

DOI: 10.35882/jeeemi.v6i4.467 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Electronics, Electromedical Engineering, and Medical Informatics, online erschienen am 16.09.2024. Verfasst von Nor Aida und 4 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Differentiated thyroid cancer is the most common type of thyroid cancer; the types in this category are papillary, follicular, and hurthel cell carcinoma. Up to 20% of DTCs will experience recurrence, although this figure reduces to 5% in low-risk patients. There is still little research on thyroid cancer prediction using a machine learning approach, especially the prediction recurrence of DTCs. This research aims to compare the performance of the Extreme Learning Machine and the Hidden Markov Model using SMOTE in predicting the recurrence of DTCs. The dataset used in this research is differentiated thyroid cancer recurrence from Kaggle. This research methodology comprises preprocessing, data sharing, SMOTE, ELM and HMM modeling algorithms, and evaluation. ELM with SMOTE gets the best results at a ratio of 90:10 with 35 hidden neurons that get an accuracy value of 1.00, precision 1.00, recall 1.00, and AUC 1.00. ELM modeling gets the best results at a ratio of 90:10 with 45 hidden neurons that get an accuracy value of 1.00, precision 1.00, recall 1.00, and AUC 1.00. HMM modeling gets the best value at a ratio of 70:30 with two hidden states and two iterations, which get an accuracy value of 0.8696, precision 0.8696, recall 0.7944, and AUC 0.9575. Last, HMM modeling with SMOTE gets the best results at a ratio of 60:40 with two hidden states and two iterations, with an accuracy value of 0.8696, precision of 0.8832, recall of 0.7848, and AUC of 0.9174. Based on the results of this study, it can be concluded that ELM with SMOTE gets the best performance, followed by ELM without SMOTE, HMM without SMOTE, and finally, HMM with SMOTE. The implication is that ELM with SMOTE can produce high accuracy in predicting the recurrence of DTCs.

Zitieren

APA

Aida, N., Saragih, T. H., Kartini, D., Nugroho, R. A., & Nugrahadi, D. T. (2024). The Comparison of Extreme Machine Learning and Hidden Markov Model Algorithm in Predicting The Recurrence Of Differentiated Thyroid Cancer Using SMOTE. Journal of Electronics, Electromedical Engineering, and Medical Informatics, 6(4), 429-444. https://doi.org/10.35882/jeeemi.v6i4.467

Vancouver

Aida N, Saragih TH, Kartini D, Nugroho RA, Nugrahadi DT. The Comparison of Extreme Machine Learning and Hidden Markov Model Algorithm in Predicting The Recurrence Of Differentiated Thyroid Cancer Using SMOTE. Journal of Electronics, Electromedical Engineering, and Medical Informatics. 2024;6(4):429-444. doi:10.35882/jeeemi.v6i4.467

BibTeX
@article{Aida2024,
  author = {Aida, Nor and Saragih, Triando Hamonangan and Kartini, Dwi and Nugroho, Radityo Adi and Nugrahadi, Dodon Turianto},
  title = {The Comparison of Extreme Machine Learning and Hidden Markov Model Algorithm in Predicting The Recurrence Of Differentiated Thyroid Cancer Using SMOTE},
  journal = {Journal of Electronics, Electromedical Engineering, and Medical Informatics},
  year = {2024},
  volume = {6},
  number = {4},
  pages = {429--444},
  doi = {10.35882/jeeemi.v6i4.467}
}
Kostenlos starten →