Title :
Monotonicity preserving SIRMs-connected fuzzy inference systems with a new monotonicity index: Learning and tuning
Author :
See Hung Lau ; Kai Meng Tay ; Chee Khoon Ng
Author_Institution :
Fac. of Eng., Univ. Malaysia Sarawak, Kota Samarahan, Malaysia
Abstract :
Recent research on Single Input Rule Modules (SIRMs)-connected fuzzy inference system (FIS) focuses on its monotonicity property fulfillment. The aim of this paper is to propose an alternative approach for modeling of monotonicity-preserving SIRMs-connected FIS. A new monotonicity index (MI) for approximating the monotonicity property fulfillment of an SIRMs-connected FIS is proposed. A hybrid of Harmony Search (HS), SIRMs-connected FIS, and the new MI is investigated. A proposed data-driven monotonicity-preserving SIRMs-connected FIS model with HS is then presented. The use of MI for tuning of an SIRMs-connected FIS is demonstrated too.
Keywords :
fuzzy reasoning; learning (artificial intelligence); search problems; HS; MI; SIRM-connected fuzzy inference system; data-driven monotonicity-preserving SIRMs-connected FIS model; harmony search; learning; monotonicity index; monotonicity preserving SIRMs-connected fuzzy inference systems; monotonicity property fulfillment; single input rule modules; tuning; Computational modeling; Fuzzy logic; Indexes; Linear programming; Mathematical model; Tuning; Vectors; data-driven; harmony search; learning; monotonicity index; single input rule modules connected fuzzy inference system; tuning;
Conference_Titel :
Fuzzy Systems (FUZZ), 2013 IEEE International Conference on
Conference_Location :
Hyderabad
Print_ISBN :
978-1-4799-0020-6
DOI :
10.1109/FUZZ-IEEE.2013.6622355
