Published Paper Details:

- Liver disease has significantly grown recently, and in many countries, it is now one of the most dangerous diseases. One of the most active and crucial organs, the liver is in charge of producing nutrients, metabolizing them, and safeguarding a non-toxic bodily environment. Its ability to remove toxins from your blood is one of its most crucial characteristics. People with liver disease are becoming more and more common as a consequence of heavy alcohol use, drugs use, polluted gas inhalation, opioids, contaminated food, and unhealthy habits. Globally, liver disease has a significant death rate. Weight loss may be some sign of liver disease. Eyes and skin that seem yellowish and leg and ankle swelling are some other signs of liver disease. To prevent the loss of lives, it is essential to identify these disorders as soon as possible. Using techniques for classification with machine learning in the healthcare sector is one of the solutions for liver disease prediction. We face a lot of data in the healthcare industry, which is one of the challenges in analyzing and studying the target condition. Choosing features that are more important than other characteristics is a difficulty in the field of disease prediction. To enhance the performance of the most accurate models, feature selection of subsets is done. This study compares models to predict liver disease, including Bayesian networks, MLP, SVM, Random Forest and Particle Swarm Optimization. The goal is to find specific features. The PSO model excels the other features in terms of the criteria for specificity, sensitivity, and accuracy. Early identification of liver diseases can save the loss of human life.