Dynamic Evolving Neuro-Fuzzy Inference System (DENFIS): On-line learning and Application for Time-Series Prediction

This paper introduces a new type of fuzzy inference systems, denoted as DENFIS (dynamic evolving neural-fuzzy system), for adaptive on-line learning, and its application for dynamic time series prediction. DENFIS evolve through incremental, hybrid (supervised/unsupervised), learning and accommodate new input data, including new features, new classes, etc. through local element tuning. New fuzzy rules are created and updated during the operation of the system. At each time moment the output of DENFIS is calculated through a fuzzy inference system based on m-most activated fuzzy rules which are dynamically chosen from a fuzzy rule set. An approach is proposed for a dynamic creation of a firstorder Takagi-Sugeno type fuzzy rule set for the DENFIS model. The fuzzy rules can be inserted into DENFIS before, or during its learning process, and the rules can also be extracted from DENFIS during, or after its learning process. An evolving clustering method (ECM), which is employed in the DENFIS model, is also introduced. It is demonstrated that DENFIS can effectively learn complex temporal sequences in an adaptive way and outperform some existing models.