Multi-label classification method and system based on correlation enhanced feature learning

The invention discloses a multi-label classification method and system based on correlation enhanced feature learning, and relates to the technical field of label classification, and the method comprises the steps: inputting an extended data set into a preset neural network for processing, and determining the input of the last layer of the preset neural network as a target data set; calculating a second similarity coefficient between the covariance matrixes of every two target data sets, reconstructing a training target set by using the second similarity coefficients, and inputting the training target set into a preset neural network to train the training target set; determining the preset neural network meeting the training ending condition as a multi-label classification network; the relations between the data features, between the labels and the data features and between the labels are captured, so that the correlation between the labels and the data features in the training data set can be reflected; and