One of the main problems in representing information in the form of nonsystematic logic is the lack of flexibility, which leads to potential overfitting. Although nonsystematic logic improves the representation of the conventional k Satisfiability, the formulations of the first, second, and third-order logical structures are very predictable. This paper proposed a novel higher-order logical structure, named G-Type Random k Satisfiability, by capitalizing the new random feature of the first, second, and third-order clauses. The proposed logic was implemented into the Discrete Hopfield Neural Network as a symbolic logical rule. The proposed logic in Discrete Hopfield Neural Networks was evaluated using different parameter settings, such as different orders of clauses, different proportions between positive and negative literals, relaxation, and differing numbers of learning trials. Each evaluation utilized various performance metrics, such as learning error, testing error, weight error, energy analysis, and similarity analysis. In addition, the flexibility of the proposed logic was compared with current state-of-the-art logic rules. Based on the simulation, the proposed logic was reported to be more flexible, and produced higher solution diversity.

No abstract

No abstract

No abstract

Choosing the best attribute from a dataset is a crucial step in effective logic mining since it has the greatest impact on improving the performance of the induced logic. This can be achieved by removing any irrelevant attributes that could become a logical rule. Numerous strategies are available in the literature to address this issue. However, these approaches only consider low-order logical rules, which limit the logical connection in the clause. Even though some methods produce excellent performance metrics, incorporating optimal higher-order logical rules into logic mining is challenging due to the large number of attributes involved. Furthermore, suboptimal logical rules are trained on an ineffective discrete Hopfield neural network, which leads to suboptimal induced logic. In this paper, we propose higher-order logic mining incorporating a log-linear analysis during the pre-processing phase, the multi-unit 3-satisfiability-based reverse analysis with a log-linear approach. The proposed logic mining also integrates a multi-unit discrete Hopfield neural network to ensure that each 3-satisfiability logic is learned separately. In this context, our proposed logic mining employs three unique optimization layers to improve the final induced logic. Extensive experiments are conducted on 15 real-life datasets from various fields of study. The experimental results demonstrated that our proposed logic mining method outperforms state-of-the-art methods in terms of widely used performance metrics.