Prediction Performance of Liver Disease based on GA with ML

Complex optimization issues are increasingly being solved by combining machine learning (ML) with genetic algorithms (GA). In high-dimensional, non-convex spaces, this hybrid method (ML-GA) is very helpful for locating global optima. The LD (Liver Disease) datasets are converted into an image-like input for the ML architecture by using both as information gains for feature selection optimization. By using a GA for better prediction performance in LD detection, this will get around the drawbacks of conventional hyper-parameter optimization methods. The sophisticated GA-based machine learning model significantly improved accuracy by outperforming conventional techniques. For both binary and multiclass LD prediction tasks, the optimized model produced a promising accuracy range using machine learning algorithms, including Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT), Logistic Regression (LR) and Random Forest (RF).