Feature gene screening and diagnosis of breast cancer based on the minimum classification error rate criterion

Breast cancer is currently the most common type of malignant cancer in women worldwide, accounting for 31 % of cancers in women, and has been on the rise in terms of both morbidity and mortality. Feature gene screening is essential for diagnosing, prognosis, and timely treatment. This study proposed a breast cancer feature genes screening method based on the minimum classification error rate criterion for accurate breast cancer diagnosis. Firstly, the overlapping area between the two distribution curves of cancer and normal gene expression data, namely, the statistically minimum classification error rate was calculated, and the breast cancer feature genes were then pre-screened from The Cancer Genome Atlas (TCGA) data with the minimum classification error rate criterion. Secondly, the feature genes were further screened based on the Weighted Gene Co-expression Network Analysis (WGCNA) and Protein-Protein Interaction (PPI) network analysis, and the Bayesian network for diagnosing breast cancer was constructed based on the screened genes. Finally, the effectiveness of the genetic screening method was validated using TCGA data within the Bayesian network diagnostic model. Experimental results showed that the method proposed in this paper had an accuracy of 96.67%, precision of 100%, recall of 93.1%, and F1 score of 0.9643, which were improved by 5%, 7.14%, 3.44%, and 5.7% compared to the conventional cancer gene screening methods with differential expression analysis.