Prediction of Liver Cirrhosis Using Classification Algorithms

Abstract

The liver is known as the largest internal organ of the body and is well known for its unique property of regeneration. One of the most common diseases of the liver is liver cirrhosis. Liver cirrhosis is one of the most widespread chronic diseases. It is characterized by the gradual replacement of the liver tissue with the scarring tissue. Liver cirrhosis, being asymptomatic, makes it difficult to identify the disease and diagnose it. Due to this factor, there is no cure for liver cirrhosis, but rather to prevent the spread of cirrhosis and reverse its effects. The main objective of this case study is to employ machine learning techniques to predict the probability of being affected by this disease. Various machine learning algorithms are employed for classifying whether a patient will develop liver cirrhosis or not [5-7]. By using various classification algorithms, the highest accuracy was achieved through Logistic Regression, XG Boost, and KNN. The accuracy was 81% for all three cases. Logistic regression, with an execution time of 0.094 seconds, outperformed the other two models. This shows that logistic regression provides a significantly more accurate prediction of liver cirrhosis disease.