Fuzzy Rule Based (FRB) and Neuro-fuzzy systems are commonly used as a basis for intelligent systems due to their transparent and simple human interpretable structure. One of the crucial steps in designing FRB and neuro-fuzzy systems is to innovate the rule base. Data clustering is one of the approaches that have been applied extensively to automatically generate rules from input-output data. The goal of this paper is to critically review some of the most commonly used as well as recently developed clustering techniques, emphasizing their use in rule base generation. The paper explores the shift from offline clustering techniques to online and finally to evolving techniques that originated due to the current demand of adaptive systems.