Publication

Abstract : In this research, the EHR data of clinical records are analyzed using the RNN model and the CT and MRI scan images of the tumour are analyzed using the CNN-based models. The models show different accuracy levels; InceptionV3+RNN achieves the best of them all by 97.7%. VGG16+RNN almost predicts the brain tumour at an accuracy of 96.2%. The precision, recall, and F1-score metrics also confirmed the robustness of these models, as evidenced by their high capability to ascertain the classifications of the malignant or non-malignant grade of tumours. Confusion matrices show the models' abilities to predict, showing the results of the prediction capability of each model over tumour grades. From the result and analysis, InceptionV3+RNN is performing better than all other ML models. It has a greater precision, recall, and f1-score value, which means it is a good classifier for tumours grades. These results are meaningful for practical applications, especially in clinical practice. The proposed approach can be implemented in real-time to predict the cancer grades with the highest level of accuracy for better treatment.