SEMA-package: Streaming EM Approximation algorithm-multilevel model: fixed...
In L-Ippel/SEMA: Streaming Expectation Maximization Approximation

Description The SEMA package includes the following functions Author(s)

The SEMA package allows one to fit online, or in a data stream, multilevel models including fixed and random intercepts and slopes. The equation numbers refer to the equations in Raudenbush and Bryk hierarchical linear models, 2nd edition.

The `SEMA` package includes the following functions

build_dataset: creates a multilevel, long-form, data set.
build_theta_j: creates all objects at the unit level.
create_theta_main: allows users to include own start values.
is.sema: checks whether list is sema class.
sema_fit_df: Fits multilevel models, input a data.frame, returns model parameters, units' parameters, and optional, stores the model parameters during the updating process.
sema_fit_one: Fits multilevel models, input one data vector, returns updated model and unit parameters.
sema_fit_set: Fits multilevel models, input one data vector, returns updated model parameters and a list with all units' parameters, one of them updated.
predict: predicts outcome based on the data of the covariates.
ranef: returns a data frame with coefficients of the random effects.
store_fixed_coef: stores the estimates of the fixed effects coefficients.
store_random_var: stores the estimates of the variance of the random effects.
store_resid_var: stores the estimate of residual variance coefficients
summary_sema: returns a list with the number of units, number of observations, estimates of the fixed effects coefficients, estimates of the variance of the random effects and the estimate of the residual variance