lncRNA-disease association prediction based on optimizing measures of multi-graph regularized matrix factorization

Abstract

In this paper, we propose a novel lncRNA-disease association prediction algorithm based on optimizing measures of multi-graph regularized matrix factorization (OM-MGRMF). The method first calculates the semantic similarity of diseases, the functional similarity of lncRNAs, and the Gaussian similarity of both. It then constructs a new lncRNA-disease association matrix by using the K-nearest-neighbor (KNN) algorithm. Finally, the objective function is constructed through the utilization of ranking measures and multi-graph regularization constraints. This objective function is iteratively optimized by an adaptive gradient descent algorithm. The experimental results of OM-MGRMF outperform those of classical methods in both K-fold cross-validation.