Radial basis function bipolar fuzzy neural network

Abstract

This paper deals with radial basis function neural network on bipolar fuzzy sets. Bipolar Gaussian positive and negative membership function acts as activation function in this network. An algorithm is developed based on Radial basis function bipolar fuzzy neural network. Further, an example is given based on the technique of injection molding process using thermoplastic raw materials, to select the highly durable plastic material product. Four thermoplastic materials such as Polyethylene terephthalate, Polyvinyl chloride, Polypropylene, Polystyrene with physical composition such as oxygen barrier, rigidity, moisture barrier, resistance to sunlight are examined. The plastic processing technology is compared with radial basis function bipolar fuzzy neural network. Considering the feeding section as the input layer, the processing section as hidden layer and ejecting section as output layer, the durability of the thermoplastic materials are estimated. It is found that Polyvinyl chloride bottles are more enduring than the other three thermoplastic material bottles.