Liver Cirrhosis Stage Classification Using Ensemble Machine Learning Techniques

Liver Cirrhosis Stage Classification Using Ensemble Machine Learning Techniques

Jenisha Blessy R1, Ms.Nishan A H2

 1Student / Information Technology, Francis Xavier Engineering College, Tirunelveli, India

jenishablessyr.ug.21.it@francisxavier.ac.in

2Assistant Professor / Information Technology, Francis Xavier Engineering College, Tirunelveli, India

nishan@francisxavier.ac.in

Abstract :

Liver cirrhosis is a chronic and dangerous condition that, if not detected early, can result in serious and often fatal complications. Early detection allows for improved treatment strategies and intervention plans to slow the progression of cirrhosis and enhance patient outcomes. However, most patients cannot afford early detection due to the high costs, invasive procedures, and lengthy time requirements of traditional detection methods like biopsies and radiological examinations. To solve this problem, we have created a highly advanced system on the basis of machine learning algorithms to forecast the development of liver cirrhosis. Rather than relying on expensive and time-consuming tests, our model employs key indicators of health such as bilirubin, albumin, platelet count, and liver enzyme activity to determine the degree of cirrhosis. Our model is derived from a Voting Classifier that utilizes the strengths of CatBoost, XGBoost, and LightGBM to make predictions. On the basis of real-world patient data, the system is able to predict the stages correctly as No Cirrhosis, Early Cirrhosis, Moderate Cirrhosis, and Severe Cirrhosis. In order to make the technology user-friendly, we designed a web application using Flask that allows health professionals to input patient information and get real-time predictions. The web interface is convenient and simple to use and provides results for classification along with personalized medical recommendations. The user can also be given a detailed PDF report with the diagnosis, input medical parameters, and recommendations for future action in the form of further medical evaluation.In order to ensure the system works as expected in real life, the model is comprehensively tested against measures like precision, recall, accuracy, and F1-score. The accuracy of misclassification is checked with a confusion matrix, which helps improve the model's performance.ease diagnosis, and improves patient care through early treatment. This system's machine learning and user-friendly interface make it easier, faster, and more accurate for a wider range of users to identify liver cirrhosis. By enabling early interventions, this tool improves patient care, expedites physician decision-making, and facilitates early detection. It can be applied to preventive care, telemedicine, and hospitals.

Keywords - Liver Cirrhosis Classification, Machine Learning, Voting Classifier, CatBoost, XGBoost, LightGBM, Predictive Analytics, Feature Selection, Disease Progression Analysis, Preventive Healthcare.