Fuzzy Rule-Based Classification Systems (FRBCS) have emerged as a powerful model in machine learning and decision-making, with its capability of handling imprecision and providing understandable models in diverse applications. The generated rules can sometimes be difficult to interpret in FRBCS, when dealing with a large number of rules. Additionally, a high number of attributes requires a larger set of rules to accurately model them which leads to an increased computational complexity. It is also crucial to consider the rules length and generate rules with short premises in order to improve the system’s interpretability. Researchers use various techniques to reduce the number of attributes as well as the number of rules. Within this article we are interested in a technique based on attributes regrouping. We study a particular method called SIFRA, which we relied on to generate local fuzzy rules in the simple fuzzy grid with the aim of improving the system performance and maintaining its interpretability and explainability.