Optimizing Lung Cancer Prediction: A Hybrid Model Integrating Hyperband And Xgboost For Enhanced Feature Extraction From Signal-Producing Images

Abstract

Lung cancer prediction has encountered challenges due to the slow learning rates of conventional models. This research introduces a hybrid model combining Hyperband optimization with the XGBoost algorithm, specifically tailored for feature extraction from signal-producing images, such as CT scans and MRI. The integration of Hyperband facilitates rapid hyperparameter tuning, while XGBoost contributes a robust gradient-boosting framework. The focus is on harnessing these advanced methodologies to improve the extraction and processing of complex features from medical images, thereby elevating predictive accuracy. The comparative analysis of this hybrid model against traditional lung cancer prediction models highlights its effectiveness in overcoming the slow learning rate issue. Results indicate not only a substantial enhancement in prediction accuracy but also a marked increase in learning efficiency, positioning this model as a valuable asset in early lung cancer detection and aiding in clinical decision-making.