This paper proposes a hybrid method for fault diagnosis of rotary kiln based on a Fuzzy
Levenberg-Marquardt (FLM) and Genetic Algorithm (GA). The huge datasets with great
number of fuzzy rules will not only significantly slow the learning phase, but also confuse the
classifier due to unusable or redundant rules. To overcome this problem, a selection method
is proposed for identifying the relevant rules and removing the irrelevant, redundant or noisy
ones. In other words, the goal of FLM-GA method is to improve classification accuracy with …