Рақамли технологияларнинг назарий ва амалий масалалари Том 8 № 2 (2025) · с. 102-107
Comparative Analysis of Machine Learning Algorithms for Breat Detection Using MRT Images Logistic Regression SVM Random Forest and CNN
Туракулов, Ш.Х.
Аннотация
Breast cancer remains a critical global health concern, demanding advanced diagnostic tools for early detection. This study assesses the effectiveness of multiple machine learning algorithms—Logistic Regression, Support Vector Machine (SVM), Random Forest, and Convolutional Neural Networks (CNN)—in analyzing Magnetic Resonance Imaging (MRI) scans for breast cancer detection. Each model classifies MRI images as malignant or benign, capitalizing on unique strengths: Logistic Regression’s simplicity, SVM’s robustness, Random Forest’s ensemble method, and CNN’s deep feature extraction. The methodology includes preprocessing MRI data, extracting features like texture and intensity, and training on a labeled dataset. Performance is evaluated using accuracy, precision, recall, and F1-score. Results show CNN leading with 96.0% accuracy and 98.6% recall (false negatives reduced from 30 to 10), followed by Logistic Regression at 94.5% accuracy (recall 97.1%), Random Forest at 93.0% (recall 94.0%), and SVM at 92.5% (recall 93.5%). Logistic Regression, despite lower accuracy, excels in computational efficiency and interpretability, ideal for resource-limited settings. Feature selection bolsters model robustness across all algorithms. This study highlights the balance between complexity and clinical utility, with CNN offering superior accuracy and Logistic Regression ensuring accessibility. Future research may explore hybrid models to enhance breast cancer diagnosis across diverse healthcare settings.
Рак молочной железыМРТмашинное обучениелогистическая регрессияметод опорных векторов (SVM)случайный лессверточные нейронные сети (CNN)извлечение признаковдиагностическая точностьклиническая применимость.
Источник метаданных: OAI-PMH архив журнала · Sindex не хранит полный текст, а даёт ссылку на источник.