The Current and Future Use of Ridge Regression for Prediction in Quantitative Genetics

Table 5

Bayes information criterion (BIC) and the proportion of explained variance () of the model for the gain in predictive accuracy of RR relative to RSR and the model for the performance of RSR, over different combinations of levels of the factors sample size, number of SNPs, fraction of causal SNPs, and heritability. Lowest BIC printed bold.

Outcome	Regressors (number of regressors)	Number of observations		BIC
	Intercept (1)	1,254,168	0.0%
	Regressors in Table 4 (5)	1,254,168	4.7%
	& 2-way interactions (15)	1,254,168	8.4%
	& 3-way interactions (35)	1,254,168	12.4%
	Intercept (1)	1,239,721	0.0%
	Regressors in Table 4 (5)	1,239,721	48.6%
	& 2-way interactions (15)	1,239,721	56.3%
	& 3-way interactions (35)	1,239,721	60.9%