Ovarian cancer prognosis prediction method based on cross-modal view association discovery network

The invention discloses an ovarian cancer prognosis prediction method based on a cross-modal view association discovery network, and the method comprises the steps: introducing a random forest and LASSO regression combined feature selection method RLASSO, removing redundant and noisy features, and fully selecting genes related to ovarian cancer prognosis; clinical features are introduced, and are respectively integrated with mRNA expression, DNA methylation, miRNA expression and copy number variation; learning advanced feature representation of specific omics data in parallel by adopting a multi-modal deep neural network, and performing ovarian cancer initial prognosis prediction; and constructing a discovery tensor for an initial prediction result by using a cross-modal view association network, and exploring cross correlation of cross omics in a space to realize final ovarian cancer prognosis prediction. The method can effectively solve the problem that the existing method neglects the difference and cross