Data dynamic classification method and system based on graph structure

Abstract

The invention discloses a data dynamic classification system based on a graph structure. The system comprises the following units: a graph data input unit used for receiving dynamically input graph structure data; a data and model storage unit is used for storing input data and storing classification model parameters at the current moment; a graph continuous learning unit used for dynamically learning classification and updating a model according to input graph data, wherein the learning graph generation unit is used for generating a learning graph under specified space limitation, and a graphlearning unit is used for updating classification model parameters according to the learning graph; and a classification prediction unit used for providing a classification prediction result for label-free data. The invention also discloses the data dynamic classification method based on the graph structure, and the system and method can dynamically learn the continuously updated graph structuredata, and can be used for continuously updated graph databases of social networks, citation networks, protein networks and the like.

Description

technical field [0001] The invention relates to the field of graph data processing, in particular to a graph structure-based data dynamic classification method and system. Background technique [0002] The graph data structure represents the complex relationship between data through its own topology. Data classification based on graph structure is a widely used classification technology. With the help of the association between data, it is possible to learn the label information of a small amount of data and predict the remaining data. classification results. However, the graph structure processed by this type of technology is often static. In real applications, there are dynamic changes in graph data, such as data changes and new data additions. Static graph data learning for such data will consume a lot of calculations every time. resources, so it poses a challenge to the existing technology, and requires an efficient classification technology and method that can handle d...

AI Technical Summary

Problems solved by technology

However, the graph structure processed by this type of technology is often static. In real applications, there are dynamic changes in graph data, such as data changes and new data additions. Static graph data learning for such data will consume a lot of calculations every time. resources, so it poses a challenge to the existing technology, and requires an efficient classification technology and method that can handle dynamic graph structure data

This technology not only needs to maintain a low computational learning efficiency for the continuously input graph data, but also needs to maintain an effective classification of all input data at all times.

Images