Acquisition method for negative label weights, terminal equipment and medium

The invention is applied to the technical field of the Internet, and provides an acquisition method for negative label weights, terminal equipment and a medium. The method comprises the steps that ifit is detected that customer relation network topology is updated, customer sample data corresponding to multiple nodes respectively is acquired; a neural network model is built and trained on the basis of the customer sample data; portrait characteristics of various to-be-audited customers in a service approval system are input to the neural network model to output weight values of all the to-be-audited customers on each negative label type; and according to the average value of the various weight values on the same negative label type, the label weight of the negative label type is computed.According to the method, automatic updating of the label weights is achieved, and the operation complexity is reduced; meanwhile, the problem that the label weight analysis dimension is single is avoided, and therefore the co