MISSIM: Improved miRNA-Disease Association Prediction Model Based on Chaos Game Representation and Broad Learning System

Abstract

MicroRNAs (miRNAs) play critical roles in the development and progression of various diseases. However, traditional experimental approaches are difficult to detect potential human miRNA-disease associations from the vast amount of biological data. Therefore, computational techniques could be of significant value. In this work, we proposed a miRNA sequence similarity calculation model (MISSIM) to large-scale predict miRNA-disease associations by combined Chaos Game Representation (CGR) with Broad Learning System (BLS). In the five-cross-validation experiment, MISSIM achieved ACC of 0.8424 on the HMDD.