Optimizing Liver Disease Detection Through Combining Genetic Evolutionary Algorithm and Linear Discriminant Analysis (LDA)

Abstract

Liver diseases such as cirrhosis, hepatocarcinoma and fatty liver disease are global health problems with high morbidity and mortality. Early detection is crucial but is often hampered by the limitations of conventional methods in analyzing medical images and laboratory results. Machine learning and artificial intelligence technologies, particularly Genetic Evolutionary Algorithm (GA) and Linear Discriminant Analysis (LDA), offer opportunities to improve diagnosis accuracy. This research explores the combination of GA and LDA to improve liver disease detection using the ILPD (Indian Liver Patient Dataset) dataset from the UCI Machine Learning Repository. This study aims to optimize feature selection and classification to improve detection accuracy. The research method includes the use of GA for feature selection and LDA for dimensionality reduction and classification. Tests were conducted on various parameters such as the number of generations, population size, and the combination of crossover and mutation rates in the genetic algorithm. The test results show that the best parameter combination (generation 400, population size 40, crossover rate 0.9, and mutation rate 0.1) results in an Average Forecast Error Rate (AFER) value of 0.0345%, which indicates that the developed detection model is highly accurate. This study shows that the combination of GA and LDA can improve the effectiveness of liver disease detection compared to conventional methods, with potential practical applications in clinical diagnosis systems.