Model optimization method and device, computer equipment and storage medium

Abstract

Embodiments of the invention disclose a model optimization method and device, computer equipment and a storage medium. The method comprises the steps of obtaining target map data; performing semi-supervised learning on the first model by adopting each node feature and target label information in the target graph data to obtain a second model; calling a second model to carry out label prediction on each label-free node to obtain a prediction label of each label-free node and the confidence coefficient of each prediction label; on the basis of the confidence of each prediction label, selecting one or more difficult samples from the N non-label nodes, wherein the difficult samples refer to the non-label nodes corresponding to the prediction labels with the confidence meeting a preset condition; and performing model optimization on the second model by adopting one or more difficult samples to obtain the target model, so that the performance of the graph processing model can be improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a model optimization method, device, computer equipment and storage medium. Background technique [0002] With the development of science and technology, the graph processing model emerges as the times require. The so-called graph processing model is a model that can predict labels for each node in the graph data. At present, when using a sample graph to optimize the model training of the graph processing model, it is usually necessary to obtain the node labels of each node in the sample graph first, so that the graph processing model can be supervised learning based on the node labels of each node, so as to realize Optimization of graph processing models. The research shows that in the existing model optimization method, labelers label each node in advance, resulting in a waste of labor costs; and the entire training optimization logic is relatively simple, which ma...

AI Technical Summary

Problems solved by technology

Research shows that in the existing model optimization method, labelers label each node in advance, resulting in a waste of labor costs; and the entire training optimization logic is relatively simple, which makes the performance of the graph processing model poor

Images