Deep Learning Techniques in the Cancer-Related Medical Domain: A Transfer Deep Learning Ensemble Model for Lung Cancer Prediction

Abstract

Problem: Cancer is regarded as one of the world's deadliest diseases. Machine learning and its new branch (deep learning) algorithms can facilitate the way of dealing with cancer, especially in the field of cancer prevention and detection. Traditional ways of analyzing cancer data have their limits, and cancer data is growing quickly. This makes it possible for deep learning to move forward with its powerful abilities to analyze and process cancer data. Aims: In the current study, a deep-learning medical support system for the prediction of lung cancer is presented. Methods: The study uses three different deep learning models (EfficientNetB3, ResNet50 and ResNet101) with the transfer learning concept. The three models are trained using a CT lung cancer dataset consisting of 1000 images and four different classes. The data augmentation process is applied to prevent overfitting, increase the size of the data, and enhance the training process. Score-level fusion and ensemble learning are also used to get the best performance and solve the low accuracy problem. All models were evaluated using accuracy, precision, recall, and the F1-score. Results: Experiments show the high performance of the ensemble model with 99.44% accuracy, which is better than all of the current state-of-the art methodologies. Conclusion: The current study's findings demonstrate the high accuracy and robustness of the proposed ensemble transfer deep learning using various transfer learning models