estimates network connection probability by network mixing of Li and Le (2021) with B-fold averaging.
network.mixing.Bfold(A,B=10,rho = 0.1,max.K=15,dcsbm=TRUE,usvt=TRUE,ns=FALSE,
lsm=FALSE,lsm.k=4)a list of
estimated probability matrix by linear mixing
estimated probability matrix by NNL mixing
estimated probability matrix by exponential mixing
estimated probability matrix by ECV mixing (only one nonzero)
the names of all individual models, in the same order as the weights
adjacency matrix
number of random replications to average over
hold-out proportion as validation entries. Only effective when index is NULL.
the maximum number of blocks used for the block model approximation (see details).
whether to include the DCSBM as components, up to max.K. By default, the method will include it.
whether to include the USVT as a component. By default, the method will include it.
whether to include the neighborhood smoothing as a component.
whether to include the gradient estimator of the latent space model as a component.
the dimension of the latent space. Only effective if lsm is TRUE.
Tianxi Li and Can M. Le
Maintainer: Tianxi Li <tianxili@virginia.edu>
This is essentially the same procedure as the network.mixing, but repeat it B times and return the average. Use with cautious. Though it can make the estimate more stable, the procedure would increase the computational cost by a factor of B. Based on our limited experience, this is usually not a great trade-off as the improvement might be marginal.
T. Li and C. M. Le, Network Estimation by Mixing: Adaptivity and More. arXiv preprint arXiv:2106.02803, 2021.
network.mixing
dt <- RDPG.Gen(n=200,K=3,directed=TRUE)
A <- dt$A
fit <- network.mixing.Bfold(A,B=2)
Run the code above in your browser using DataLab