Machine Learning-Based Liver Disease Prediction: Enhancing Diagnosis and Prognosis

Recently, liver diseases have emerged as a significant health concern globally, with a rising number of cases attributed to factors such as alcohol consumption, exposure to harmful substances, dietary choices, and certain medications. To address this issue, liver patient datasets were analyzed to develop classification models aimed at predicting liver disorders and reducing the burden on healthcare providers. In this study, we employ machine learning algorithms to conduct a comprehensive assessment of chronic liver disorders, defined by conditions persisting for at least six months. Utilizing patient health indicators, including positive and negative disease status markers, we apply various classification techniques to analyze liver disease prevalence and present the results in a confusion matrix. This research introduces the concept of Liver Disease Prediction (LDP) and evaluates multiple clustering and classification algorithms to improve prediction accuracy. Key algorithms used include KMeans clustering, Logistic Regression (LR), Decision Tree (DT), and Support Vector Machine (SVM). The combination of KMeans with LR and DT demonstrated high accuracy, achieving an accuracy rate of 99.9852 %, while the combination of KMeans and SVM attained 99.9705%. These results suggest that machine learning models, particularly hybrid models, can achieve near-perfect accuracy, enhancing diagnostic support for liver disease prediction.