Research on new fuzzy deep learning model and its construction technology

Abstract

The application of deep learning in adaptively extracting corresponding feature expressions from a large number of unbalanced data sets for classification has become a hot topic of research and discussion at home and abroad in recent years. The purpose of this paper is to study the new model of fuzzy deep learning and its construction technology. A vehicle detection algorithm based on fuzzy deep belief network is proposed. Deep belief fuzzy networks can gain the ability to integrate prior knowledge by introducing fuzzy set theory into deep belief networks. It is a deep framework that combines the power of abstract restricted Boltzmann machines with the power of fuzzy set classification. Constrained Boltzmann functions can achieve fast data dimensionality reduction, and fuzzy sets can improve the classification accuracy of deep learning frameworks based on membership functions for each class. The experimental results on the wine dataset show that the detection algorithm based on fuzzy deep belief network proposed in this paper can classify faster and more accurately.