Extract Classification Rules from COVID-19 Datasets using Multi-Objective Advanced Genetic Algorithm

Authors:
Publisher website
Google Scholar
Abstract
The rapid proliferation of Coronavirus Disease 2019 (COVID-19) has resulted in a substantial escalation of suspected and confirmed cases globally, posing a significant threat to public health, economic stability, and international relations This underscores the urgent need for alternative diagnostic strategies that leverage artificial intelligence to enhance pandemic response efficacy. This research endeavors to augment predictive accuracy through the Multi-Objective Advanced Genetic Algorithm (MoAGA), which automatically generates classification rules from datasets. To this end, the Multi-Objective Single Genetic Algorithm (MoSGA) was enhanced by incorporating a distributed population strategy alongside an optimized migration procedure. Furthermore, the proposed method was refined through the integration of two fuzzy control systems, which dynamically adjust crossover and mutation rates, culminating in the development of the Multi-Objective Fuzzy Genetic Algorithm (MoFGA). This innovative approach optimizes both population diversity and fitness. The methodology's efficacy was validated against COVID-19-related datasets, with performance assessed through metrics such as prediction accuracy, maximum fitness, and average fitness. Results indicated that the MoAGA achieved a prediction accuracy of 99.30%, with a maximum fitness of 0.88 and an average fitness of 0.86. This novel methodology demonstrates a superior ability to uncover rule sets with markedly higher predictive precision than existing techniques, enhancing the potential for optimal global solutions.

This publication has 33 references indexed in Scilit: