The present invention provides a method for identifying the association between piRNA and
disease based on self-
supervised learning, comprising the following steps: step 1, constructing a
benchmark data set for the study of the association between piRNA and
disease; step 2, calculating multi-dimensional piRNA similarity and
disease similarity; step 3, creating a
feature matrix based on multiple views; step 4, constructing an identification model for the association between piRNA and disease based on self-
supervised learning, and obtaining an association recognition
score; step 5, using the recognition
score obtained in step S4, analyzing and quantifying in detail whether there is an edge relationship between piRNA and disease, so as to identify the association between piRNA and disease. The present invention is used to identify the association between piRNA and disease, and it constructs a reliable
data set based on three public databases, and calculates similarity to create a multi-view
feature matrix; in model identification,
singular value decomposition is used to optimize the relationship between nodes, a combined augmentation strategy is used to retain key information, and a KAN network is used to make subsequent classification results more reliable and credible.