Use Harris Hawks Optimization (HHO) Algorithm based on Artificial Neural Network for liver disease diagnosis

Abstract

This research paper presents a new method for diagnosing liver disease using the Harris Hawks Optimization (HHO) algorithm in combination with an Artificial Neural Network (ANN). The HHO algorithm enhances the ANN's performance in liver disease classification by refining its parameters. Clinical, laboratory, and demographic data are collected from hepatitis patients and individuals without hepatic illness. The dataset is processed to handle missing values, outliers, and normalization. The HHO algorithm optimizes the weights and biases of the ANN, facilitating the identification of relevant features for accurate diagnosis. The trained ANN model is evaluated using various performance metrics, demonstrating its effectiveness in diagnosing liver disease. The HHO algorithm efficiently searches the entire search space, enhancing the ANN's ability to learn complex patterns and make accurate predictions. Evaluation metrics indicate that the optimized ANN model outperforms traditional machine learning methods, showcasing its potential as a reliable diagnostic tool. Interpretability metrics, such as feature importance and saliency maps, provide insights into the key elements of diagnosis. The proposed approach shows high diagnostic accuracy and interpretability, implying the potential for a stable decision-aid system in clinical practice. Early detection and timely intervention enabled by this method can lead to increased patient safety rates and optimized resource allocation.