Method for predicting correlation between microorganisms and drugs based on double attention map convolution

The invention discloses a method and device for predicting correlation between microorganisms and drugs based on double attention map convolution, and the method comprises the steps: constructing a microorganism and drug correlation network subgraph based on an obtained known microorganism-drug correlation relation; learning an association prediction map representation of the microorganism and drug association network subgraph by using an attention map convolutional neural network; and introducing the correlation prediction map representation into an attention pooling layer so as to extract information under maximized node representation and obtain a microorganism-drug prediction model. According to the method for predicting microorganism and drug association based on double attention graph convolution, context information of a specific sub-graph is automatically captured by applying an attention graph convolution network and an attention pooling layer so as to perform expression feature learning, so that a p