Predicting Lung Cancer Death Rates with a Novel Deep Learning Approach to Assist Medical Decision-Making

Abstract

Introduction: The aim of this proposed approach Elevating Naïve Bayes Optimization with Improved Convolutional Neural Networks (ENBO-ICNN), is to predict lung cancer (LC) death rates while considering the complexities and ethical concerns inherent in the function of deep learning (DL) algorithms in the healthcare domain.
Objective: To address the limitations of data quality, disease complexity and the evolving nature of research in forecasting LC death rates by implementing the ENBO-ICNN approach.
Method: The study gathered data from the NCRP dataset, preprocessing using Min-Max normalization and employing Principal Component Analysis (PCA) for feature extraction to develop the ENBO-ICNN approach for predicting LC mortality rates.
Result: The efficacy of the ENBO-ICNN technique is demonstrated with its enhanced performance measures, which include significant recall, accuracy, precision, and F1-score in improving LC death rate predictions compared to existing methods.
Conclusion: The proposed ENBO-ICNN approach signifies a significant advancement in predicting LC death rates, offering valuable insights into the complexities of the disease and addressing critical ethical considerations associated with DL algorithms in healthcare.