The prevalence of liver disease is rising, and because the liver can still function normally even when it is partially destroyed, its early stages might be difficult to detect. To increase the patient's chances of survival, it is essential to recognize liver disease as soon as possible. Expert physicians are needed to identify liver illness using a variety of examination techniques, albeit this does not ensure a correct diagnosis. Even experienced doctors have trouble telling apart disease from common symptoms. A misdiagnosis of the condition occurs because the bulk of the symptoms are similar to those of other fever-related illnesses. A computer-aided diagnosis is necessary for accurate liver disease prediction and for managing enormous amounts of data because other diseases predominate and make liver disease difficult to identify. In this study, we are evaluating the diagnostic evaluation of several machine learning algorithms on liver disease datasets. To combine dimensionality reduction and optimization approaches with the chosen classification algorithms to assess result variation. To conduct a performance study on classification algorithms and to choose the most appropriate classification models for liver disorders.