A method and system for filling key information of target nodes based on an association network

Abstract

The invention discloses a method and system for filling key information of target nodes based on an association network, and belongs to the technical fields of data mining, machine learning and graph theory. The problem of low accuracy of key information of the filled target node in the prior art is solved. According to the application scenario, the present invention establishes a large number of node relational networks; based on each relational network, obtains the associated network of the target node with key information, and integrates the associated network into an object node, label, associated node, associated node weight and attribute vector Data structure, multiple three-dimensional sampling of the data structure based on the improved random forest method to obtain a subset of multiple training decision trees, given multiple decision trees for training, and integrating after training to obtain the final model; based on the target nodes to be filled The associated nodes are predicted through the final model, and after the prediction, multiple results are weighted and averaged to obtain the final filling information. The present invention fills in the key information of the target node based on the association network.

Description

technical field [0001] A method and system for filling key information of a target node based on an association network, used for filling key information of a target node based on an association network, belonging to the technical fields of data mining, machine learning, and graph theory. Background technique [0002] In many scenarios, there is a need to predict the key information of the target when there is insufficient target information. Specific scenarios include the fields of financial credit, e-commerce recommendation, health assessment, and other fields. [0003] Scenario 1: In the field of financial credit, how to conduct credit evaluation for credit white account access. The credit white account itself does not have enough basic credit information for financial institutions to evaluate its repayment willingness and repayment ability. At this time, the relevant information of the target node's close relatives (that is, the adjacent network nodes) can be used as th...

AI Technical Summary

Problems solved by technology

[0015] In view of the problems of the above research, the purpose of the present invention is to provide a method and system for filling key information of target nodes based on an associated network, to solve the problem in the prior art that (1) when filling key information of target nodes, it is necessary to rely on the target node itself The characteristics of the target node, in the case that the target node has no relevant features, it is impossible to fill the key information of the target node; (2) the accuracy of the key information of the target node is low; (3) the problem of occupying a lot of resources

Images