Fit NEMoE model

This function fit NEMoE model using EM algorithm.

fitNEMoE(
  NEMoE_obj,
  beta_init = NULL,
  gamma_init = NULL,
  num_restart = 5,
  restart_it = 10,
  BPPARAM = BiocParallel::SerialParam(),
  ...
)

Arguments

NEMoE_obj	a NEMoE object contain data and parameters for fitting.
beta_init	initial values of parameters in experts network.
gamma_init	initial value of parameters in gating network.
num_restart	Number of restart times.
restart_it	Maximial number of iterations during restart. the BiocParallel package is used. Default is SerialParam().
BPPARAM	A BiocParallelParam class object from.
...	other parameters can be passed to fitNEMoE. See createParameterList

Value

a NEMoE object contain fitted result including parameters beta a list contain coefficients of experts network in all levels. gamma a matrix of coefficients in gating network. r_i an array of inherit partition result in EM algorithm. pen_fac estimated penalty factor for lambda1 (adapt = T). The (balanced) lambda1 in NEMoE is lambda1*pen_fac latent estimated latent classes.

See also

NEMoE_buildFromList, NEMoE_buildFromPhyloseq, createParameterList

Examples

data(NEMoE_example)
NEMoE_example
#> Numer of samples:  168 
#> Microbiome number of variables: 5, 16, 24, 36, 101, 
#> Nutrition dim: 168 27 
#> Number of latent class K:  2 
#> NEMoE output names (5): beta gamma r_i pen_fac LL
#> param names (14): lambda1 lambda2 ... verbose early_stop
#> cross validation param names (9): g1 lambda2_seq ... itmax_cv itmax_fit
NEMoE_example <- fitNEMoE(NEMoE_example, num_restart = 0)
#> Fitting NEMoE.... 
#> it: 0, PLL0:-1032.93
#> it: 1, PLL:-521.229
#> it: 2, PLL:-511.879
#> it: 3, PLL:-508.635
#> it: 4, PLL:-506.598
#> it: 5, PLL:-505.111
#> it: 6, PLL:-503.808
#> it: 7, PLL:-502.643
#> it: 8, PLL:-501.587
#> it: 9, PLL:-500.677
#> it: 10, PLL:-499.79
#> it: 11, PLL:-498.912
#> it: 12, PLL:-498.135
#> it: 13, PLL:-497.429
#> it: 14, PLL:-496.733
#> it: 15, PLL:-495.999
#> it: 16, PLL:-495.196
#> it: 17, PLL:-494.288
#> it: 18, PLL:-493.227
#> it: 19, PLL:-491.953
#> it: 20, PLL:-490.378
#> it: 21, PLL:-488.453
#> it: 22, PLL:-486.091
#> it: 23, PLL:-483.381
#> it: 24, PLL:-480.661
#> it: 25, PLL:-478.076
#> it: 26, PLL:-475.773
#> it: 27, PLL:-473.822
#> it: 28, PLL:-472.299
#> it: 29, PLL:-471.153
#> it: 30, PLL:-470.36
#> it: 31, PLL:-469.868
#> it: 32, PLL:-469.542
#> it: 33, PLL:-469.326
#> it: 34, PLL:-469.179
#> it: 35, PLL:-469.077
#> it: 36, PLL:-469.008
#> it: 37, PLL:-468.962
#> it: 38, PLL:-468.93
#> it: 39, PLL:-468.906
#> it: 40, PLL:-468.888
#> it: 41, PLL:-468.872
#> it: 42, PLL:-468.859
#> it: 43, PLL:-468.848
#> it: 44, PLL:-468.838
#> it: 45, PLL:-468.829
#> it: 46, PLL:-468.821
#> it: 47, PLL:-468.813
#> it: 48, PLL:-468.806
#> it: 49, PLL:-468.797
#> it: 50, PLL:-468.789
#> it: 51, PLL:-468.778
#> it: 52, PLL:-468.766
#> it: 53, PLL:-468.752
#> it: 54, PLL:-468.733
#> it: 55, PLL:-468.714
#> it: 56, PLL:-468.696
#> it: 57, PLL:-468.679
#> it: 58, PLL:-468.663
#> it: 59, PLL:-468.648
#> it: 60, PLL:-468.632
#> it: 61, PLL:-468.617
#> it: 62, PLL:-468.602
#> it: 63, PLL:-468.588
#> it: 64, PLL:-468.573
#> it: 65, PLL:-468.56
#> it: 66, PLL:-468.549
#> it: 67, PLL:-468.539
#> it: 68, PLL:-468.531
#> it: 69, PLL:-468.525
#> it: 70, PLL:-468.52
#> it: 71, PLL:-468.516
#> it: 72, PLL:-468.514
#> it: 73, PLL:-468.513
#> it: 74, PLL:-468.513
#> it: 75, PLL:-468.513
#> it: 76, PLL:-468.513
#> it: 77, PLL:-468.513
#> it: 78, PLL:-468.513