The invention discloses an intestinal
flora and
obesity personalized diagnosis method based on a graph
attention network, and relates to the technical field of
obesity diagnosis, and the method comprises the following steps: collecting
gene data, building a graph structure-based model for the data to predict
obesity risk and provide personalized health suggestions, each piece of
microbiome data is expressed as a weighted
undirected graph, a
microbiome graph is constructed, the association strength between
microbiome groups is calculated, and the microbiome data is expressed as a graph structure. According to the method, the graph structure model considering the relevance between the microbiome groups is constructed, the self-adaptive graph convolutional network and the heterogeneous graph
attention network are utilized, multi-
modal information interaction between the microbiome and clinical data is deeply mined, the
processing capacity of high-dimensional and complex microbiome data is improved, and the multi-
modal information interaction between the microbiome and clinical data is improved. The accurate prediction of obesity risks and the provision of personalized health suggestions are realized, the prediction accuracy and generalization ability of the model are significantly improved, and a more effective personalized obesity diagnosis effect is achieved.