30-03-2017, 02:56 PM
In the field of artificial intelligence, neuro-fuzzy refers to combinations of artificial neural networks and fuzzy logic. Neuro-diffuse hybridization results in a hybrid intelligent system that synergizes these two techniques by combining the human reasoning style of fuzzy systems with the learning and connectionist structure of neural networks. Neurofuzzy hybridization is widely referred to as a blurred neural network (FNN) or neurofuzzy system (NFS) in the literature. The neurofuzzy system (the most popular term used hereafter) incorporates the human reasoning style of fuzzy systems through the use of fuzzy sets and a linguistic model consisting of a set of IF-THEN diffuse rules. The main strength of neuro-diffuse systems is that they are universal approximators with the ability to request interpretable IF-THEN rules.
The strength of neuro-diffuse systems involves two contradictory requirements in diffuse modeling: interpretability versus accuracy. In practice, one of the two properties prevails. The field of neurofuzzy research in diffuse modeling is divided into two areas: fuzzy linguistic modeling that focuses on interpretability, mainly the Mamdani model; And accurate fuzzy modeling that focuses on accuracy, mainly the Takagi-Sugeno-Kang (TSK) model.
Although generally assumed to be the realization of a fuzzy system through connectionist networks, this term is also used to describe some other configurations including:
• Derive diffused rules from trained RBF networks.
• Fuzzy logic based on the tuning of training parameters of the neural network.
• Fuzzy logic criteria to increase the size of the network.
• Performing the fuzzy membership function through clustering algorithms in unsupervised learning in SOM and neural networks.
• Representation of fuzzification, fuzzy inference and defuzzification through multi-chamber connection-connection networks.
It should be noted that the interpretability of Mamdani-like neurofuzes can be lost. To improve the interpretability of neuro-diffuse systems, certain measures must be taken, in which important aspects of the interpretability of neuro-diffuse systems are also discussed.
The strength of neuro-diffuse systems involves two contradictory requirements in diffuse modeling: interpretability versus accuracy. In practice, one of the two properties prevails. The field of neurofuzzy research in diffuse modeling is divided into two areas: fuzzy linguistic modeling that focuses on interpretability, mainly the Mamdani model; And accurate fuzzy modeling that focuses on accuracy, mainly the Takagi-Sugeno-Kang (TSK) model.
Although generally assumed to be the realization of a fuzzy system through connectionist networks, this term is also used to describe some other configurations including:
• Derive diffused rules from trained RBF networks.
• Fuzzy logic based on the tuning of training parameters of the neural network.
• Fuzzy logic criteria to increase the size of the network.
• Performing the fuzzy membership function through clustering algorithms in unsupervised learning in SOM and neural networks.
• Representation of fuzzification, fuzzy inference and defuzzification through multi-chamber connection-connection networks.
It should be noted that the interpretability of Mamdani-like neurofuzes can be lost. To improve the interpretability of neuro-diffuse systems, certain measures must be taken, in which important aspects of the interpretability of neuro-diffuse systems are also discussed.
