POPFNN-AAR(S): a pseudo outer-product based fuzzy neural network

Abstract

A novel fuzzy neural network, the pseudo outer-product-based fuzzy neural network using the singleton fuzzifier together with the approximate analogical reasoning schema, is proposed in this paper. The network is referred to as the singleton fuzzifier POPFNN-AARS, the singleton fuzzifier POPFNN-AARS employs the approximate analogical reasoning schema (AARS) instead of the commonly used truth value restriction (TVR) method. This makes the structure and learning algorithms of the singleton fuzzifier POPFNN-AARS simple and conceptually clearer than those of the POPFNN-TVR model. Different similarity measures (SM) and modification functions (FM) for AARS are investigated. The structures and learning algorithms of the proposed singleton fuzzifer POPFNN-AARS are presented. Several sets of real-life data are used to test the performance of the singleton fuzzifier POPFNN-AARS and their experimental results are presented for detailed discussion.

A novel self-organizing fuzzy rule-based system for modelling traffic flow behaviour

“…The results obtained indicate that the modelling ability of POPFNN-TVR is comparable to and often better than of feed-forward neural networks using conventional back-propagation learning (FFBP) (Rumelhart, Hinton, & Williams, 1986). It also performs better on this problem that similar architectures (Ang, Quek, & Pasquier, 2003;Quek & Zhou, 1999), that have been thoroughly benchmarked against other methods. More importantly, POPFNN-TVR offers the added advantage over FFBP and other pure connectionist approaches of providing a set of explicit knowledge rules extracted from the training data.…”

“…The rule identification method used in POPFNN is the onepass Lazy Pseudo Outer Product (LazyPOP) learning algorithm (Quek & Zhou, 1999;Zhou & Quek, 1996), which is closely related to Hebbian's learning law. The use of POPFNN as the fuzzy-neural network for traffic flow prediction has the advantages to be efficient, convenient, highly intuitive, and easier to understand than other rule identification algorithms.…”

Hierarchically Clustered Adaptive Quantization CMAC and Its Learning Convergence

“…Currently, these research endeavours are actively underway at the C2i [55]. The C2i lab undertakes intense research in the study and development of advanced brain-inspired learning memory architectures [74]- [78] for the modeling of complex, dynamic, and nonlinear systems. These techniques have been successfully applied to numerous novel applications such as automated driving [58], signature forgery detection [79], gear control for the continuous variable transmission (CVT) system in an automobile [80], fingerprint verification [81], bank failure classification and early warning system (EWS) [82], computational finance [83], [84], as well as in the biomedical engineering domain [85], [86].…”

Maximizing winning trades using a novel RSPOP fuzzy neural network intelligent stock trading system

“…Pseudo Outer-Product based Fuzzy Neural Networks (POPFNN) is a family of neuro-fuzzy systems that is based on the linguistic fuzzy model [16]. Three members of POPFNN exists in the literature, namely: POPFNN-CRI(S) [25] which is based on commonly accepted fuzzy Compositional Rule of Inference, POPFNN-TVR [26] which is based on Truth Value Restriction, and POPFNN-AARS(S) [27] which is based on the Approximate Analogical Reasoning Scheme. The POPFNN architecture is a five-layer neural network as shown in Fig.…”