Construction and analysis of a joint diagnosis model of random forest and artificial neural network for heart failure.

Heart failure is a global health problem that affects approximately 26 million people worldwide. As conventional diagnostic techniques for heart failure have been in practice with various limitations, it is necessary to develop novel diagnostic models to supplement existing methods. With advances and improvements in gene sequencing technology in recent years, more heart failure-related genes have been identified. Using existing gene expression data in the Gene Expression Omnibus (GEO) database, we screened differentially expressed genes (DEGs) of heart failure and identified six key genes (HMOX2, SERPINA3, LCN6, CSDC2, FREM1, and ZMAT1) by random forest classifier. Of these genes, CSDC2, FREM1, and ZMAT1 have never been associated with heart failure. We also successfully constructed a new diagnostic model of heart failure using an artificial neural network and verified its diagnostic efficacy in public datasets.