Beyin manyetik rezonans görüntülerinde çoklu skleroz lezyonları için harrıs hawks optimizasyonu (HHO) algoritmasına dayalı veri madenciliği yaklaşımı

Segmentation approaches, the majority of which are based on the combination of data mining and machine learning techniques, can be used to extract damaged sections of brain tissues. Clustering algorithms, such as the fuzzy C-means (FCM) cluster algorithm, are effective for segmentation since they are accurate yet not unduly sensitive to visually noise. As a consequence, while the optimum cluster center selection might impact segmentation, the FCM approach is acceptable for detecting Multiple Sclerosis. It's complicated to choose them since it's an NP-hard problem. Harris Hawks optimization (HHO) was utilized to determine the optimal cluster center for both segmentation and FCM methods. Other common methods that aren't as exact as HHO include the genetic algorithm and particle swarm optimization. In the suggested procedure, each membership matrix must either be a hawk or an HHO. To reduce Multiple Sclerosis clustering mistakes, the next step is to construct hawk populations or membership vectors, with the best one chosen to locate the optimal cluster centers. The suggested technique outperformed FCM clustering, as well as other methods like the support vector machine, k-NN algorithm, and hybrid data mining methods, in accuracy testing on a variety of brain MRIs.