Compound Classification Method Based on Graph Neural Network

Abstract

The invention discloses a compound classification method based on a graph neural network, which is used to solve the problems that the existing classification methods ignore the structural information in the compound classification, resulting in inaccurate classification results and low classification efficiency. The steps of the present invention are: (1) constructing two graph neural networks; (2) generating a training set with category labels and a training set without category labels; (3) training two graph neural networks; (4) Classify compounds that do not contain class labels. The present invention builds and trains two graph neural networks, which can better capture the structural information contained in the compound, and adopts pre-training, collaborative training and self-training methods so that the present invention has a shorter time when dealing with compounds with complex information. Processing time and high compound classification accuracy.

Description

technical field [0001] The invention belongs to the technical field of physics, and further relates to a compound classification method based on a graph neural network in the technical field of image classification. According to the structural features and molecular node attributes of the modeled compound graph, the present invention can extract the attribute information of the compound molecule from the graph structure and node attributes through the graph neural network, and classify according to the information, such as judging whether a certain compound is Antibiotic molecule compound. Background technique [0002] Compounds, as real-life non-Euclidean data, can be naturally represented by graph data structures, which are typically used to represent a set of objects (i.e., nodes) and their relationships (i.e., edges between nodes). In conventional compound classification techniques, a series of physical or chemical experiments are usually required to determine the chara...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the shortcomings of the above-mentioned prior art, and propose a compound classification method based on graph neural network, which is used to solve the problem that the existing classification method ignores the structural information in the compound classification, resulting in inaccurate classification results and low classification efficiency. low problem

Images