Diagnosis of Liver Diseases Using Machine Learning Algorithms and their Prediction Using Logistic Regression and ANN

Globally, liver diseases affect a large number of population and they are the 11th leading cause of death. It is tedious and time-consuming for trained experts to identify patients with liver disease through liver biopsy and MRI checks. Liver disease can be difficult to identify in the early stages due to mild symptoms. As a result, symptoms emerge when it is too late. A timely diagnosis can potentially save a person's life. Machine Learning can detect early signs of these diseases, even when they are not readily apparent to even experienced doctors. Therefore, the objective of this work is to assess how well various machine-learning algorithms perform. Support Vector Machine, Logistic Regression, Naïve Bayes, Conventional Neural Network, and Artificial Neural Network were the five algorithms we employed in this study. The performance of these techniques was evaluated using a variety of measuring metrics, such as accuracy, f-1 score, precision, and recall. The accuracy was 65.26%, 73.65%, 70.05%, 97.5%, and 93.75% for SVM, LR, NB, CNN, and ANN. The analysis's findings indicate that CNN achieved the greatest accuracy. Furthermore, the current study focused on the use of clinical data and imaging scans to predict liver disease.