Main Article Content
This study focuses on the application of formal logic systems to real-world problem-solving, specifically in the classification of the COVID-19 Surveillance Data Set (CSDS). The research introduces the integration of a random three satisfiability problem of Boolean logic into a Hopfield Neural Network (HNN) to obtain an optimal representation of Random kSatisfiability for CSDS classification. The primary goal is to utilize the optimization capabilities of the Lyapunov energy function in the HNN to extract logical relationships and identify significant features contributing to COVID-19 detection. The CSDS used in this study is sourced from the reputable UCI dataset, and the HNN's energy minimization mechanism is employed for logical mining. Computational simulations are performed with varying numbers of clauses to validate the efficacy of the proposed model in training the CSDS for classification purposes. The results showcase the efficiency and robustness of employing reverse analysis using k-satisfiability in conjunction with a Hopfield Neural Network. This approach successfully extracts dominant features related to the logical framework underlying the CSDS. By combining formal logic systems with the power of neural networks, this research offers insights into the correlation between logical rules and COVID-19 detection. The findings contribute to our understanding of how the HNN can effectively learn and classify data, opening avenues for enhanced classification techniques in the healthcare sector and other domains.
This work is licensed under a Creative Commons Attribution 4.0 International License.