Analysis of Hepatitis Patient Data using Binary Artificial Algae Algorithm based on K-Nearest Neighbor

Abstract

Introduction: The timely diagnosis and prediction of diseases are among the main issues in medical sciences. The use of decision-making systems to discover the underlying knowledge in the disease information package and patient records is one of the most effective ways of diagnosing and preventing disease. This study aimed to design a medical decision system that can detect hepatitis. Materials & Methods: This study was conducted based on a descriptive-analytic design. Its dataset contains 155 records with 19 features in the University of California-Irvine machine learning database. This study utilized the Binary Artificial Algae Algorithm (BAAA) for Feature Selection (FS). Moreover, K-Nearest Neighbor (KNN) was used to classify hepatitis into two healthy and unhealthy classes. In total, 80% of the data was employed for training, and the remaining (20%) was used for testing. Furthermore, Precision, Recall, F-measure, and Accuracy were utilized to evaluate the model. Findings: According to the results, the accuracy of the proposed model was estimated at 96.45%. After selecting the features with the BAAA, the percentage of the accuracy reached 98.36% in the best situation. In the proposed model with 300 repetitions, the Precision, Recall, F-Measure, and error rate were 96.23%, 96.74%, 96.48%, and 3.55%, respectively. Discussion & Conclusions: Hepatitis is one of the most common diseases among females and males. A timely diagnosis of this disease not only reduces the costs but also increases the chance of successful treatment. In this study, the disease was diagnosed using the hybrid method, and a high accuracy level was obtained in disease diagnosis by FS.

Keywords

• binary artificial algae algorithm
• feature selection
• hepatitis disease diagnosis
• k-nearest neighbor
• medical decision making system