LncRNA and disease association prediction method and device based on graph auto-encoder, and computer storage medium

The invention discloses an LncRNA and disease association prediction method and device based on a graph auto-encoder and a computer storage medium, and belongs to prediction of an LncRNA and disease association relationship. The problems that according to an existing correlation prediction method, the nonlinear relation or the linear relation between the disease and the LncRNA cannot be effectively captured, the prediction efficiency is low, and the time cost and the economic cost are high are solved. The method comprises the following steps: S1, obtaining known association information among LncRNA, MiRNA and a disease; s2, obtaining semantic similarity of the disease; s3, obtaining the functional similarity of the LncRNA and the functional similarity of the MiRNA; s4, adopting a matrix splicing mode to obtain a heterogeneous network; s5, obtaining a plurality of feature sub-graphs; s6, obtaining a feature vector corresponding to the LncRNA and the disease node pair; and S7, obtaining a correlation degree pre