Fig. 1
From: A fall risk prediction model based on the CHARLS database for older individuals in China
Selection of demographic and clinical characteristics using the LASSO regression model. A Coefficient profiles were generated from the log (lambda) series, and nonzero coefficients were generated from the optimal lambda. B The optimal parameter (lambda) in the LASSO model was selected by ten-fold cross-validation using the minimum criterion. Partial likelihood deviation (binomial deviation) curves are plotted relative to log (lambda). A virtual vertical line is drawn at the optimum using one SE of the minimum criterion (1-SE criterion)