Unbalanced data classification method based on cyclic consistent generative adversarial network

The invention discloses an unbalanced data classification method based on a cyclic consistent generative adversarial network, and mainly solves the problems that in the prior art, a classification model is sensitive to noise data, inter-class overlapping exists in synthetic samples of a data layer method, and the diversity of synthetic sample features is deficient. The method comprises the following implementation steps: 1) preprocessing an original data set; 2) constructing a twin data pair set according to the preprocessed data; 3) designing a cyclic consistent generative adversarial network in which two groups of discriminators and generators are arranged; 4) performing iterative training on the network model by using the data in the twin data pair set, and synthesizing target minority class sample data; and 5) enhancing target minority class sample data to an original data set to obtain a balanced data set, and training a basic classifier by using the data set to complete classification. The method can re