Closed tmaklin closed 1 year ago
Running the lowrank algorithm on linear coefficients currently requires projecting the coefficients via the relationship BX and supplying the projected coefficients and the design matrix via
BX
cpprate -x design_matrix_file.csv -f projected_linear_coefficients_file.csv
A more intuitive way to call the lowrank algorithm with linear coefficients would be something like
cpprate -x design_matrix_file.csv --beta-draws linear_coefficients_file.csv
where the program handles projecting the linear coefficients and runs the lowrank algorithm if the --fullrank toggle is not used.
--fullrank
Running the lowrank algorithm on linear coefficients currently requires projecting the coefficients via the relationship
BXand supplying the projected coefficients and the design matrix viaA more intuitive way to call the lowrank algorithm with linear coefficients would be something like
where the program handles projecting the linear coefficients and runs the lowrank algorithm if the
--fullranktoggle is not used.