Machine Learning Based Liver Cancer Disease Prediction System Using Improved Extreme Gradient Boosting Algorithm

Diagnosing and treating liver cancer poses a challenge, for healthcare systems globally. However current liver cancer prediction models struggle with predicting the disease in cancerous areas. This study introduces an improved XGBoost model for predicting liver cancer disease that requires time compared to existing models. The model is. Tested using the Indian Liver Cancer Patient Dataset (ILPD). Data imputation is carried out using the imputer class from the sklearn library followed by implementing Univariate feature selection to identify features. These selected features are then input into the improved Extreme Gradient Boosting (XGBoost) algorithm to predict liver cancer disease. Evaluation of this model includes metrics such as Accuracy, Precision, Recall and F1 Score, with numerical validation results showing 94.67%, 99.25%, 92.57% and 95.65% respectively. The new approach is being evaluated against models, in use such as the Prediction of chronic liver cancer disease patients using integrated projection based statistical feature extraction with machine learning algorithms (PCLDP SFXML) and Multi Source Transfer Learning, via Multi Kernel Support Vector Machine Plus for B Mode Ultrasound Based Computer Aided Diagnosis of Liver Cancers (MKL NHSVM).