The use of Ontology in Diagnosing LC: A Detailed Framework Discussion

Accurate diagnosis and classification of lung cancer is essential for effective surgical management and drug development. The complexity of this effort comes from the difficulty of analyzing CT (computed tomography) images. Our proposed solution reduces false positives, segments jumps, and classifies images as normal and abnormal using deep learning techniques using fuzzy-K-means clustering The system has two main methods: Feature extraction, preprocessing, and classification phase: Texture features extracted after preprocessing input CT image. Crow search optimization methods are used to select features, and artificial neural networks are used to classify images as correct or incorrect. The ontology and fuzzy K-means method is then used to analyze the abnormal images for tumor classification. To reduce false positives, a support vector machine (SVM) classifier and ontology is used in this step. Impressive outcomes are obtained by the suggested approach, which successfully lowers false positives while achieving a $\mathbf{9 6 \%}$ accuracy rate, 100% specificity, and $99 \%$ sensitivity. Experimental results show improved sensitivity, specificity, and accuracy when compared to previous studies. In addition, our system achieves a notable reduction in computing time by finding a compromise between efficiency and efficacy. Our approach demonstrates a straightforward yet effective combination of deep learning techniques with fuzzy K-means clustering.