Predictive Modeling for Cirrhosis Diagnosis: A Machine Learning Exploration

Cirrhosis manifests as a condition wherein healthy liver tissue is gradually replaced by scar tissue, resulting in a decline in liver function. This ailment frequently correlates with autoimmune disorders, non-alcoholic fatty liver disease, excessive alcohol intake, and persistent liver conditions such as hepatitis B or C. As cirrhosis progresses, it can cause liver failure and various complications such as ascites, portal hypertension, hepatic encephalopathy, and an increased risk of liver cancer. Since the liver plays a crucial role in filtering blood and performing essential functions, individuals with acute liver failure are more prone to infections, particularly those affecting the circulatory, respiratory, and urinary systems. Kidney failure is a common consequence of liver failure, especially in cases of acetaminophen overdose, as this medication can damage both the liver and kidneys. In the realm of healthcare, the integration of machine learning has enabled the early prediction of liver cirrhosis using historical datasets. The aim of this research is to create a machine learning model that utilizes artificial intelligence to forecast the occurrence of liver cirrhosis using supplied data. By employing hyperparameter tuning techniques to enhance accuracy, this research employs machine learning techniques like logistic regression and Gradient Boosted Classifier using XGBoost. Utilizing GridSearch within XGBoost, this research work achieves an impressive accuracy rate of 98%, accompanied by precision, recall, and F1-scores of 96%, 97%, and 96%, respectively. These findings hold promise for the advancement of more sophisticated and accurate liver cirrhosis prediction models, thereby benefiting researchers in the field.