<![CDATA[The purpose of this paper is to suggest an integrated system that combines the classification of medical images and a recommendation model for the choice of the best hospitals for cancer treatment according to the analysis of the brain MRI. For the classification of brain MRI images for tumor detection, we used EfficientNet-B3 as the base model. To evaluate its performance, we contrasted it with two other popular deep learning architectures, namely Vision Transformer (ViT) and ResNet101. Having identified malignancies by image classification, we used the outputs as inputs for a collaborative filtering based recommendation system. This system was constructed utilizing neural network embeddings and hidden vectors to learn the correlations between tumor types and hospital treatment characteristics. Personalized hospital suggestions were generated by calculating the similarity between the embeddings using the dot product. In addition, comparison evaluations with a traditional K-Nearest Neighbors (K-NN) method are performed to evaluate the performance gaps. The classification results revealed that EfficientNet-B3 has better accuracy when compared with ViT and ResNet101 and it is a good choice for medical picture analysis. The neural collaborative filtering model shows superior accuracy and suitability of the hospital selection during the recommendation phase compared to K-NN. The results of this study demonstrate that the combination of deep learning and intelligent recommendation systems may successfully provide excellent diagnosis in healthcare. The proposed framework may improve decision making and may provide individualized recommendations of hospitals for cancer treatment. A case study is offered in Vietnam to recommend hospitals.]]>
Copyrights © 2026
