ANFIS Construction With Sparse Data via Group Rule Interpolation

A major assumption for constructing an effective adaptive-network-based fuzzy inference system (ANFIS) is that sufficient training data are available. However, in many real-world applications, this may not hold, thereby requiring alternative approaches. In light of observation, article focuses on automated construction ANFISs effort to enhance the potential Takagi-Sugeno regression models situations where only limited particular, proposed approach works by interpolating a group rules certain given domain with assistance existing its neighboring domains. The process involves number computational mechanisms, including rule dictionary which created extracting from ANFISs; interpolated exploiting local linear embedding algorithm build intermediate ANFIS; and fine-tuning method refines resulting ANFIS. experimental evaluation both synthetic datasets reported, demonstrating when facing shortage situations, helps significantly improve performance original ANFIS modeling mechanism.