Liver cirrhosis prediction using a machine learning approach

Abstract

Liver Cirrhosis Prediction is a crucial area of research aimed at improving the accuracy and effectiveness of identifying liver cirrhosis cases. This study explores the performance of three classification algorithms, namely Naïve Bayes, Random Forest, and Ada Boost, in predicting liver cirrhosis. The experimental results demonstrate high accuracy rates for the Naïve Bayes (97.61%) and Random Forest (98.80%) classifiers, indicating their effectiveness in classifying liver cirrhosis cases. The Naïve Bayes classifier exhibits an Ill-balanced performance with precision, recall, and f1-score values of 93, 98, and 95, respectively. The Random Forest classifier surpasses the other algorithms, achieving superior precision, recall, and f1-scores of 99, 92, and 94, respectively. The Ada Boost classifier achieved a low accuracy rate of 80.95% with precision, recall, and f1-score values of 67, 75, and 70, respectively. These findings highlight the potential of the Naïve Bayes and Random Forest classifiers in liver cirrhosis prediction, providing valuable insights for healthcare professionals and researchers. Future research could focus on refining the Ada Boost classifier and exploring hybrid models or advanced techniques to further enhance the accuracy and precision of liver cirrhosis prediction models. The successful prediction of liver cirrhosis can contribute to early intervention and improved patient outcomes in clinical settings.