cr.sparse.block.bsbl.BSBL_Options¶
- class cr.sparse.block.bsbl.BSBL_Options(learn_block_corr: int = 1, learn_lambda: int = 1, prune_gamma: float = 0.001, lambda_val: float = 1e-12, max_iters: int = 800, epsilon: float = 1e-08)[source]¶
Options for the BSBL algorithm
Attributes
epsilon
Solution accuracy tolerance parameter
lambda_val
User defined value for lambda [if provided by user]
learn_block_corr
Indicates whether intra block correlations are to be learned or ignored.
learn_lambda
Indicates if the variance of noise is to be estimated from data
max_iters
Maximum number of iterations for the BSBL algorithm
prune_gamma
Threshold for pruning small values of gamma_i