Method, device and equipment for identifying isomerism graph and property of molecular spatial structure

Abstract

The invention discloses a method and a device for identifying isomerism graph and property corresponding to a molecular space structure, and machine equipment. The method comprises the following steps: carrying out feature description containing a topological structure on a heterogeneous graph to generate feature information; through sampling the isomerism graph and the feature information, generating a feature vector, corresponding to a key node, on a topological structure contained in the isomerism graph; performing aggregation processing on the feature vectors to generate graph representation vectors corresponding to the isomerism graphs; and according to the graph representation vector, performing classification processing on the isomerism graph to obtain a classification prediction result. Therefore, the identification of the graphical representation corresponding to the heterogeneous graph, such as the molecular spatial structure, can be realized through the neural network, the analysis and processing performance related to the isomerism graph is effectively improved, the neural network is not limited to the classification and identification of network data any more, and the application scene is greatly expanded.

Description

technical field [0001] The present invention relates to the field of computer application technology, in particular to a method, device and machine equipment for realizing the recognition of heterogeneous graphs and the recognition of properties corresponding to molecular space structures. Background technique [0002] Various recognition based on machine learning has always been an important tool for researching and analyzing various source data in various application scenarios. For example, in application scenarios such as image and natural language processing, the large-scale classification and identification of source data can be implemented to quickly obtain the classification and prediction results related to the source data, and accelerate the function realization of the application scenario. [0003] As a classic implementation architecture of machine learning, neural networks have achieved great success in the field of classification and recognition. Many models and...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the technical problem in the related art that the identification of the graphical representation corresponding to the molecular spatial structure cannot be realized through the neural network, the present invention provides a heterogeneous graph and even the properties ...

Images