Evolving output-context fuzzy system for effective rule base

This study develops an evolving output-context fuzzy system (EOCFS) that initiates the evolving process with a single fuzzy rule in which consequent and antecedent parts cover the whole output and input domains respectively. The EOCFS further evolves the output domain and adds more fuzzy rules to achieve an effective rule base. The proposed model is a self-organizing method that can automatically identify prominent distinct data in the output domain for a new fuzzy rule. Thus, the EOCFS constantly evolves by reducing the model error. In addition, the evolving process is realized on the output domain while the self-adaptive process is achieved on the output domain and its associated input domain. The evolving termination index and uncertainty controller of the self-adaptive process are dynamically attained from past and current knowledge. Therefore, effective rule base is the balanced fuzzy model of the approximated system. To illustrate the effectiveness of the proposed algorithm, synthetic and real-world data are considered with low to high dimension inputs. Results show that the proposed EOCFS achieves better performance than the existing models with regard to accuracy and number of rules.