Fig. 6From: Federated learning algorithms for generalized mixed-effects model (GLMM) on horizontally partitioned data from distributed sourcesThe curve of test power among centralized, Laplace, and Gauss–Hermite methods. (Left) The power of the test of the Laplace method. (Middle) The power of the test of the 2-degree Gauss–Hermite method. (Right) The power of the test of the Centralized method. Power was calculated as the two-sided t-test on p values among different methodsBack to article page