TimeGMM(data, Timepoints, Q, case = "diag", tol = 1e-8)
Parameters
Timepoints - [t_1, t_2, ... t_T] # Timepoints for which the data is provided in the data object
data - {Timepoints[t]: n_txd_t size 2-D array | for t in Timepoints}
Q - Number of components
Returns
Connects all time points. Trains a GMM model on the first timepoint given in the Timepoint list. Estimates the v parameter based on the
stepwise_TimeGMM(X, Timepoints, Q, case = "diag", tol = 1e-8)
Connects two immediate time points. Trains a GMM model on timepoint t and extrapolates to timepoint t+1 to estimate the v. Models stepwise progression this way.
single_gaussian_best_fit(X, Timepoints, beta = 0.001, case = "diag", tol = 1e-8)