Skip to main content
Fig. 3 | Translational Medicine Communications

Fig. 3

From: Machine learning-selected variables associated with CD4 T cell recovery under antiretroviral therapy in very advanced HIV infection

Fig. 3

Comparison of linear models using A, Top 20 LASSO variables, and B, basal CD4 T cell counts, as predictors of net increment in CD4 T cell counts after 52 weeks on ART (“delta T CD4”). For A, regression was performed using leave-one-out cross validation, performing N linear regressions, training with N-1 samples, plotting the predictions for the sample that was unknown to the model. Adjusted R2 = 0.71960667, F = 170.383628 (DF = 65) p = 7.915957X10− 20. B, linear model with basal (week 0) CD4 T cell counts. Adjusted R2 = 0.034, F = 3.33 (DF = 65), p = 0.07

Back to article page