Around a million deaths are occurring due to liver diseases globally. There are several traditional methods to diagnose liver diseases, but they are expensive. Early prediction of liver disease would benefit all individuals prone to liver diseases by providing early treatment. As the technology is growing in health care, machine learning is significantly affecting health care for predicting conditions at early stages.
This study finds how accurate machine learning is in predicting liver disease. This present study introduces the liver disease prediction (LDP) method in predicting liver disease that can be utilised by health professionals, stakeholders, students and researchers.
Five algorithms, namely Support Vector Machine (SVM), Naïve Bayes, K-Nearest Neighbors (K-NN), Linear Discriminant Analysis (LDA), and Classification and Regression Trees (CART), are selected. The accuracy is compared to uncover the best classification method for predicting liver disease using R and Python. From the results, K-NN obtains the best accuracy with 91.7%, and the autoencoder network achieved 92.1% accuracy, which is above the acceptable level of accuracy and can be considered for liver disease prediction.