EUVIMEDEUVIMED entdecken
Crossref

Machine Learning Framework For Accurate Multiclass Detection Of Malignant Brain Tumors

H.Saisohan H.Saisohan

International Journal of Scientific Research in Engineering and Management · 2026-07 · 10 (7)

DOI: 10.55041/ijsrem66461 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Scientific Research in Engineering and Management, 2026-07. Verfasst von H.Saisohan H.Saisohan.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The study of malignant and non-malignant brain tumors is a major application of CAD systems, which are used by many healthcare professionals and radiologists to improve certainty in their medical decision making based on analysis of MRI images.In the past, traditional methods have focused mostly on binary classification cases, including identifying whether a brain image is normal or abnormal or whether brain tumors are benign or malignant.However, the precise identification of categories of malignant brain tumors is still a difficult problem that needs more effort.In this work, a multi-class classification method is proposed for the identification of several types of high-grade malignant brain tumors based on a real-world dataset that contains five classes of tumors.The proposed method is able to capture meaningful hidden data by extracting a large amount of data from several domains.To capture the most relevant features while avoiding data redundancy, a new feature selection method, Cumulative Variance Method (CVM), is applied.The extracted features are then used to build, for the first time, three different machine learning models, K-Nearest Neighbour (KNN), Multi-class Support Vector Machine (mSVM), and Neural Network (NN), and evaluate their performance.The average classification accuracy for KNN, mSVM, and NN is 88.43%, 92.50%, and 95.86%, respectively.In addition, the relative improvement in classification accuracy with the new feature selection method is about 2% for KNN, 3% for mSVM, and 4% for NN, when compared to other feature selection methods, namely, Independent Component Analysis (ICA) and Genetic Algorithm (GA).

Zitieren

APA

H.Saisohan, H. (2026). Machine Learning Framework For Accurate Multiclass Detection Of Malignant Brain Tumors. International Journal of Scientific Research in Engineering and Management, 10(7). https://doi.org/10.55041/ijsrem66461

Vancouver

H.Saisohan H. Machine Learning Framework For Accurate Multiclass Detection Of Malignant Brain Tumors. International Journal of Scientific Research in Engineering and Management. 2026;10(7). doi:10.55041/ijsrem66461

BibTeX
@article{HSaisohan2026,
  author = {H.Saisohan, H.Saisohan},
  title = {Machine Learning Framework For Accurate Multiclass Detection Of Malignant Brain Tumors},
  journal = {International Journal of Scientific Research in Engineering and Management},
  year = {2026},
  volume = {10},
  number = {7},
  doi = {10.55041/ijsrem66461}
}
Kostenlos starten →