Development and validation of the risk stratification model (4P-model). (IMAGE)
Caption
(A) Pipeline of model development using machine learning. CPAS-O, a gene subset that distinguishes PAS cases from the control group; CPAS-S, a gene subset that distinguishes increta/percreta cases from accreta cases; CPAS-Tri, a risk stratification model that distinguishes cases into increta/percreta, accreta, and controls. (B) Venn diagram of two gene subsets, CPAS-O and CPAS-S. (C) AUCs of SVM-RBF models with different numbers of features during RFE. AUC values were calculated from out-of-fold predictions in 10-fold cross-validation. (D) ROC plots for 15 gene-based 4P-model for increta/percreta. (E) Confusion matrices for the 4P-model in each cohort. Each cell shows the number of individuals belonging to this category, and the percentages in parentheses indicate the proportion of predicted labels in each group to the actual total of each group.
Credit
Runrun Hao, Beilun Wang, Bin Zhang, Yixuan Li, Qingwen Zhu, Shilin Zhu, Weiling Liu, Yihan Wu, Lina Zhang, Xueqi Bai, Shanshan Wang, Shiman Hu, Yuting Zhang, Haoyan Shi, Dong Liang, Huixia Yang, Zhonghua Shi
Usage Restrictions
Credit must be given to the creator. Only noncommercial uses of the work are permitted. No derivatives or adaptations of the work are permitted.
License
CC BY-NC-ND