DeepFCM and DeepFCMx are developed as transparent and comprehensive methodologies based on deep learning and FCM techniques for the automatic classification of two separate classification problems: coronary artery disease and non-small cell lung cancer.
The main strengths of DeepFCM are the effective employment of both image and clinical data. The generated weight matrix and the produced graph present full transparency of the model’s structure by demonstrating the interconnections among the concepts. The advantage of DeepFCMx is that it only requires image data as input and adjustments to the ranges among relationships among concepts of different classes based on experts’ knowledge.
In summary, the advancements in DeepFCM and DeepFCMx pave the way for a powerful and interpretable medical classification system. By effectively handling diverse data types and employing feature selection and deep learning techniques, DeepFCM has the potential to revolutionize the classification tasks of Coronary Artery Disease and Non-Small Cell Lung Cancer. Future research will focus on evaluating the effectiveness, scalability, and applicability of DeepFCM across different medical datasets, and problems and enhance the results of DeepFCMx.
