A fuzzy reasoning method based on weight of mixed matrix

There is a little study on interpolative reasoning method of multi-dimensional fuzzy variables under sparse rule base. Among them, a representative is approximate interpolative reasoning method (MB method for short), but it does not take into account weights of multi-dimensional variables which can influence conclusions, and it is difficult to ensure the convexity and normality of reasoning results. This study presents a multi-dimensional fuzzy reasoning method based on weight of mixed matrix. The method use mixed matrix chart to determine weights, and then produce new fuzzy reasoning rules with weights. Case study shows that this method not only improves accuracy, but also ensures convexity and normality of reasoning results.