ENHANCED LIVER CANCER DETECTION USING HYBRID CNN-XGBOOST MODEL IN MACHINE LEARNING
KALAISELVI N.1, SASIKALA R.2
Student1, Assistant professor2Department of Computer Science and Engineering
KIT – Kalaignar karunanidhi Institute of Technology, Coimbatore, India
Abstract: Since liver cancer is such an aggressive illness, improving patient outcomes requires early identification. This paper presents a novel detection technique that blends Extreme Gradient Boosting (XGBoost) with Convolutional Neural Networks (CNNs). Medical image complex feature extraction is a strong suit for CNNs, and XGBoost is a potent classifier that excels at processing high-dimensional data. The suggested hybrid model attempts to improve detection accuracy through integrating the deep feature extraction powers of CNNs with the effective classification of XGBoost. Based on experimental data, this CNN-XGBoost model achieves a noteworthy detection accuracy of 95.2%, outperforming standalone CNN and XGBoost classifiers. With improved accuracy and dependability, this development makes a substantial contribution to computer-aided diagnostic systems. This strategy works by giving medical practitioners quick and accurate diagnostic tools.
Keywords: Liver cancer, Early detection, Convolutional Neural Networks (CNN), Extreme Gradient Boosting (XGBoost), Machine learning, medical imaging, Hybrid models.