An optimized logic mining method for data processing through higher-order satisfiability representation in discrete hopfield neural network

A high performance classification tool such as logic mining has emerged as one of the future computing systems for big data processing. The collaboration between logic and neural network resulted in extracting the most suitable induced logic to represent knowledge from real-life datasets. However, there are certain limitations within the current logic mining models including a non-flexible logical structure, non-optimal computation of the best logic, and the generation of overfitting solutions. Motivated by these limitations, a novel logic mining model incorporating the non-systematic Satisfiability, namely Random 3 Satisfiability in Discrete Hopfield Neural Network is proposed as a logical structure to represent the behaviour of the dataset. The proposed logic mining models used flexible logical structures to prevent overfitting solutions and optimize synaptic weight values. A new computational approach of the best logic by considering True Positive and True Negative values of the learning system is applied in this work to preserve the significance behaviour of the dataset. Furthermore, the comparative experiments of the logic mining models by utilizing various repository real-life datasets are conducted from repositories to assess their efficiency. In accordance with the results, the proposed logic mining model dominates in all the metrics for the average rank. The average rank for each metrics are Accuracy (1.9375), Precision (1.9375), Specificity (1.8125), Mathews Correlation (1.5625), and Fowlkes Mallows Index (2.3125). Numerical results and in-depth analysis demonstrate that the proposed logic mining model consistently produces optimal induced logic that best represents the real-life dataset for all the performance metrics used in this study.

Text 124165.pdf - Published Version Restricted to Repository staff only Download (7MB) Official URL or Download Paper: https://www.sciencedirect.com/science/article/pii/...

Actions (login required)

View Item