Logical inference model based on fuzzy linear regression

Mathematical Modelling: Methods, algorithms, technologies

A new fuzzy inference approach based on IF-THEN rules and fuzzy linear regression is proposed. Approximation theorem concerning this model is formulated and proofed. An outline of producing fuzzy rules that approximate given table-based function is described. Results of experiments comparing proposed approach with Mamdani-type systems are given. Practical task of prognosis for a disease based on patient’s state that solved by proposed fuzzy system is described.