Figure 5
From: Predicting breast cancer risk using interacting genetic and demographic factors and machine learning
Combining the Group 1 (a) and Group 2 (b) demographic features with the genotyped data outperforms all other single systems in BC risk prediction on the KBCP test data. Literature SNPs denotes the system trained on 82 known BC-risk-associated SNPs, and the PRS model is derived from those 82 SNPs. The number after ± denotes the standard deviation. The high standard deviations are due to multiple subset selections.
