calculate_prec_from_corrse#

pharmpy.modeling.calculate_prec_from_corrse(corr, se)[source]#

Calculate precision matrix from a correlation matrix and standard errors

Parameters:

corr (pd.DataFrame) – Correlation matrix
se (pd.Series) – Standard errors

Returns:

pd.DataFrame – Precision matrix

Examples

>>> from pharmpy.modeling import calculate_prec_from_corrse
>>> from pharmpy.tools import load_example_modelfit_results
>>> results = load_example_modelfit_results("pheno")
>>> corr = results.correlation_matrix
>>> se = results.standard_errors
>>> corr
           POP_CL    POP_VC   COVAPGR    IIV_CL    IIV_VC     SIGMA
POP_CL   1.000000  0.008433 -0.095506 -0.387063  0.098882  0.168004
POP_VC   0.008433  1.000000 -0.357003  0.138290  0.356831  0.239295
COVAPGR -0.095506 -0.357003  1.000000 -0.095767 -0.062857  0.153034
IIV_CL  -0.387063  0.138290 -0.095767  1.000000 -0.185775  0.165104
IIV_VC   0.098882  0.356831 -0.062857 -0.185775  1.000000 -0.026388
SIGMA    0.168004  0.239295  0.153034  0.165104 -0.026388  1.000000
>>> calculate_prec_from_corrse(corr, se)
               POP_CL        POP_VC       COVAPGR         IIV_CL        IIV_VC          SIGMA
POP_CL   2.993418e+07  22260.995666  16027.840996  203633.422078 -39113.196303 -817311.952371
POP_VC   2.226100e+04   2129.850713    260.115336    -373.895598  -2799.329201   -7697.904374
COVAPGR  1.602784e+04    260.115336    187.053654     177.987259   -205.261518   -2224.519605
IIV_CL   2.036334e+05   -373.895598    177.987259    7542.266046   2472.031665  -10209.388516
IIV_VC  -3.911320e+04  -2799.329201   -205.261518    2472.031665  22348.204432    9193.183296
SIGMA   -8.173120e+05  -7697.904374  -2224.519605  -10209.388516   9193.183296  249977.511621