Soft Mathematical System to Solve Black Box Problem through Development the FARB Based on Hyperbolic and Polynomial Functions

This work attempts to design a method for computer-based mathematical model of multi-parametric datasets that their features are fully-automatic structural and parametric optimization of models. This method has different inputs and one output as a subset of components for the base function F. In addition, several functions in modeling the classified rules are used including 12 polynomials and 6 hyperbolic functions. This work combines the advantage of data mining algorithms and Fuzzy All-Permutation Rule Base (FARB). In order to evaluate the performance of the proposed method, two classifications based on association rules case studies were used in an attempt to design a dynamic mathematical method. The first case study takes the virtual rules which consist of four rules, while in the second case study actual rules are selected. According to the experimental results, not only enhanced accuracy can be achieved by the proposed method, but also it can be used for intelligent data analysis for huge and small datasets.