Skip to main content
Fig. 5 | BMC Medical Research Methodology

Fig. 5

From: The case-crossover design via penalized regression

Fig. 5

ROC curves of multivariable CLR and the lasso for reclassifying highly suspected drugs. Cut-off values of the log ORs for discriminating between highly suspected drugs versus suspected or other drugs range from 0 to 4; the sensitivity and 1-specificity of several cut-off values are highlighted. A 45-degree line of equality is added for reference

Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us