Modified Probabilistic 2 Satisfiability Reverse Analysis via Ensemble Attribute selection Operator

The validity of logic mining depends on the validity of logic rules and the logic mining model. Standard system logic mining technology has obtained a series of optimization methods. The most effective optimization methods include embedding permutation operators and optimal attribute selection, but they fail to map unbalanced data sets well. Therefore, In order to improve the efficiency of the reverse analysis model of the existing system logic mining and to fit the unbalanced data, this paper introduces a PRO2SAT logic mining model based on traditional reverse analysis. It uses the generalization attribute of different positive literal proportions of PRO2SAT and the Ensemble Attribute Selection (EAS) strategy to help extract the best logic from the data set in the correct search space. The experimental results show that the proposed PRO2SATRA logic mining scheme has significant improvement in accuracy and competitiveness compared with the standard 2SAT reverse analysis method.