Developer Reference for Intel® oneAPI Math Kernel Library for C

ID 766684
Date 3/22/2024
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

vslSSEditTask

Modifies address of an input/output parameter in the task descriptor.

Syntax

status = vslsSSEditTask(task, parameter, par_addr);

status = vsldSSEditTask(task, parameter, par_addr);

status = vsliSSEditTask(task, parameter, par_addr);

Include Files

  • mkl.h

Input Parameters

Name

Type

Description

task

VSLSSTaskPtr

Descriptor of the task

parameter

const MKL_INT

Parameter to change

par_addr

const float* for vslsSSEditTask

const double* for vsldSSEditTask

const MKL_INT* for vsliSSEditTask

Address of the new parameter

Output Parameters

Name

Type

Description

status

int

Current status of the task

Description

The vslSSEditTask routine replaces the pointer to the parameter stored in the Summary Statistics task descriptor with the par_addr pointer. If you pass the NULL pointer to the editor, no changes take place in the task and a corresponding error code is returned. See Table "Parameters of Summary Statistics Task to Be Initialized or Modified" for the predefined values of the parameter.

Use the single (vslsssedittask) or double (vsldssedittask) version of the editor, to initialize single or double precision version task parameters, respectively. Use an integer version of the editor (vslissedittask) to initialize parameters of the integer type.

Parameters of Summary Statistics Task to Be Initialized or Modified

Parameter Value

Type

Purpose

Initialization

VSL_SS_ED_DIMEN

i

Address of a variable that holds the task dimension

Required. Positive integer value.

VSL_SS_ED_OBSERV_N

i

Address of a variable that holds the number of observations

Required. Positive integer value.

VSL_SS_ED_OBSERV

d, s

Address of the observation matrix

Required. Provide the matrix containing your observations.

VSL_SS_ED_OBSERV_STORAGE

i

Address of a variable that holds the storage format for the observation matrix

Required. Provide a storage format supported by the library whenever you pass a matrix of observations.1

VSL_SS_ED_INDC

i

Address of the array of indices

Optional. Provide this array if you need to process individual components of the random vector. Set entry i of the array to one to include the ith coordinate in the analysis. Set entry i of the array to zero to exclude the ith coordinate from the analysis.

VSL_SS_ED_WEIGHTS

d, s

Address of the array of observation weights

Optional. If the observations have weights different from the default weight (one), set entries of the array to non-negative floating point values.

VSL_SS_ED_MEAN

d, s

Address of the array of means

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_2R_MOM

d, s

Address of an array of raw moments of the second order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_3R_MOM

d, s

Address of an array of raw moments of the third order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_4R_MOM

d, s

Address of an array of raw moments of the fourth order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_2C_MOM

d, s

Address of an array of central moments of the second order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first and second order.

VSL_SS_ED_3C_MOM

d, s

Address of an array of central moments of the third order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first, second, and third order.

VSL_SS_ED_4C_MOM

d, s

Address of an array of central moments of the fourth order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first, second, third, and fourth order.

VSL_SS_ED_KURTOSIS

d, s

Address of the array of kurtosis estimates

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first, second, third, and fourth order.

VSL_SS_ED_SKEWNESS

d, s

Address of the array of skewness estimates

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first, second, and third order.

VSL_SS_ED_MIN

d, s

Address of the array of minimum estimates

Optional. Set entries of array to meaningful values, such as the values of the first observation.

VSL_SS_ED_MAX

d, s

Address of the array of maximum estimates

Optional. Set entries of array to meaningful values, such as the values of the first observation.

VSL_SS_ED_VARIATION

d, s

Address of the array of variation coefficients

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array. Make sure you also provide arrays for raw moments of the first and second order.

VSL_SS_ED_COV

d, s

Address of a covariance matrix

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Make sure you also provide an array for the mean.

VSL_SS_ED_COV_STORAGE

i

Address of the variable that holds the storage format for a covariance matrix

Required. Provide a storage format supported by the library whenever you intend to compute the covariance matrix.2

VSL_SS_ED_COR

d, s

Address of a correlation matrix

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. If you initialize the matrix in non-trivial way, make sure that the main diagonal contains variance values. Also, provide an array for the mean.

VSL_SS_ED_COR_STORAGE

i

Address of the variable that holds the correlation storage format for a correlation matrix

Required. Provide a storage format supported by the library whenever you intend to compute the correlation matrix.2

VSL_SS_ED_ACCUM_WEIGHT

d, s

Address of the array of size 2 that holds the accumulated weight (sum of weights) in the first position and the sum of weights squared in the second position

Optional. Set the entries of the matrix to meaningful values (typically zero) if you intend to do progressive processing of the dataset or need the sum of weights and sum of squared weights assigned to observations.

VSL_SS_ED_QUANT_ORDER_N

i

Address of the variable that holds the number of quantile orders

Required. Positive integer value. Provide the number of quantile orders whenever you compute quantiles.

VSL_SS_ED_QUANT_ORDER

d, s

Address of the array of quantile orders

Required. Set entries of array to values from the interval (0,1). Provide this parameter whenever you compute quantiles.

VSL_SS_ED_QUANT_QUANTILES

d, s

Address of the array of quantiles

None.

VSL_SS_ED_ORDER_STATS

d, s

Address of the array of order statistics

None.

VSL_SS_ED_GROUP_INDC

i

Address of the array of group indices used in computation of a pooled covariance matrix

Required. Set entry i to integer value k if the observation belongs to group k. Values of k take values in the range [0, g-1], where g is the number of groups.

VSL_SS_ED_POOLED_COV_STORAGE

i

Address of a variable that holds the storage format for a pooled covariance matrix

Required. Provide a storage format supported by the library whenever you intend to compute pooled covariance.2

VSL_SS_ED_POOLED_MEAN

d, s

Address of an array of pooled means

None.

VSL_SS_ED_POOLED_COV

d, s

Address of pooled covariance matrices

None.

VSL_SS_ED_GROUP_COV_INDC

i

Address of an array of indices for which covariance/means should be computed

Optional. Set the kth entry of the array to 1 if you need group covariance and mean for group k; otherwise set it to zero.

VSL_SS_ED_REQ_GROUP_INDC

i

Address of an array of indices for which group estimates such as covariance or means are requested

Optional. Set the kth entry of the array to 1 if you need an estimate for group k; otherwise set it to zero.

VSL_SS_ED_GROUP_MEANS

i

Address of an array of group means

None.

VSL_SS_ED_GROUP_COV_STORAGE

d, s

Address of a variable that holds the storage format for a group covariance matrix

Required. Provide a storage format supported by the library whenever you intend to get group covariance.2

VSL_SS_ED_GROUP_COV

d, s

Address of group covariance matrices

None.

VSL_SS_ED_ROBUST_COV_STORAGE

d, s

Address of a variable that holds the storage format for a robust covariance matrix

Required. Provide a storage format supported by the library whenever you compute robust covariance2.

VSL_SS_ED_ROBUST_COV_PARAMS_N

i

Address of a variable that holds the number of algorithmic parameters of the method for robust covariance estimation

Required. Set to the number of TBS parameters, VSL_SS_TBS_PARAMS_N.

VSL_SS_ED_ROBUST_COV_PARAMS

d, s

Address of an array of parameters of the method for robust estimation of a covariance

Required. Set the entries of the array according to the description in vslSSEditRobustCovariance.

VSL_SS_ED_ROBUST_MEAN

i

Address of an array of robust means

None.

VSL_SS_ED_ROBUST_COV

d, s

Address of a robust covariance matrix

None.

VSL_SS_ED_OUTLIERS_PARAMS_N

d, s

Address of a variable that holds the number of parameters of the outlier detection method

Required. Set to the number of outlier detection parameters, VSL_SS_BACON_PARAMS_N.

VSL_SS_ED_OUTLIERS_PARAMS

i

Address of an array of algorithmic parameters for the outlier detection method

Required. Set the entries of the array according to the description in vslSSEditOutliersDetection.

VSL_SS_ED_OUTLIERS_WEIGHT

d, s

Address of an array of weights assigned to observations by the outlier detection method

None.

VSL_SS_ED_ORDER_STATS_STORAGE

d, s

Address of a variable that holds the storage format of an order statistics matrix

Required. Provide a storage format supported by the library whenever you compute a matrix of order statistics.1

VSL_SS_ED_PARTIAL_COV_IDX

i

Address of an array that encodes subcomponents of a random vector

Required. Set the entries of the array according to the description in vslSSEditPartialCovCor.

VSL_SS_ED_PARTIAL_COV

d, s

Address of a partial covariance matrix

None.

VSL_SS_ED_PARTIAL_COV_STORAGE

i

Address of a variable that holds the storage format of a partial covariance matrix

Required. Provide a storage format supported by the library whenever you compute the partial covariance.2

VSL_SS_ED_PARTIAL_COR

d, s

Address of a partial correlation matrix

None.

VSL_SS_ED_PARTIAL_COR_STORAGE

i

Address of a variable that holds the storage format for a partial correlation matrix

Required. Provide a storage format supported by the library whenever you compute the partial correlation.2

VSL_SS_ED_MI_PARAMS_N

i

Address of a variable that holds the number of algorithmic parameters for the Multiple Imputation method

Required. Set to the number of MI parameters, VSL_SS_MI_PARAMS_SIZE.

VSL_SS_ED_MI_PARAMS

d, s

Address of an array of algorithmic parameters for the Multiple Imputation method

Required. Set entries of the array according to the description in vslSSEditMissingValues.

VSL_SS_ED_MI_INIT_ESTIMATES_N

i

Address of a variable that holds the number of initial estimates for the Multiple Imputation method

Optional. Set to p+p*(p+1)/2, where p is the task dimension.

VSL_SS_ED_MI_INIT_ESTIMATES

d, s

Address of an array of initial estimates for the Multiple Imputation method

Optional. Set the values of the array according to the description in "Basic Components of the Multiple Imputation Function in Summary Statistics" in the Intel® oneAPI Math Kernel Library (oneMKL) Summary Statistics Application Notes document [SS Notes].

VSL_SS_ED_MI_SIMUL_VALS_N

i

Address of a variable that holds the number of simulated values in the Multiple Imputation method

Optional. Positive integer indicating the number of missing points in the observation matrix.

VSL_SS_ED_MI_SIMUL_VALS

d, s

Address of an array of simulated values in the Multiple Imputation method

None.

VSL_SS_ED_MI_ESTIMATES_N

i

Address of a variable that holds the number of estimates obtained as a result of the Multiple Imputation method

Optional. Positive integer number defined according to the description in "Basic Components of the Multiple Imputation Function in Summary Statistics" in the Intel® oneAPI Math Kernel Library (oneMKL) Summary Statistics Application Notes document [SS Notes].

VSL_SS_ED_MI_ESTIMATES

d, s

Address of an array of estimates obtained as a result of the Multiple Imputation method

None.

VSL_SS_ED_MI_PRIOR_N

i

Address of a variable that holds the number of prior parameters for the Multiple Imputation method

Optional. If you pass a user-defined array of prior parameters, set this parameter to (p2+3*p+4)/2, where p is the task dimension.

VSL_SS_ED_MI_PRIOR

d, s

Address of an array of prior parameters for the Multiple Imputation method

Optional. Set entries of the array of prior parameters according to the description in "Basic Components of the Multiple Imputation Function in Summary Statistics" in the Intel® oneAPI Math Kernel Library (oneMKL) Summary Statistics Application Notes document [SS Notes].

VSL_SS_ED_PARAMTR_COR

d, s

Address of a parameterized correlation matrix

None.

VSL_SS_ED_PARAMTR_COR_STORAGE

i

Address of a variable that holds the storage format of a parameterized correlation matrix

Required. Provide a storage format supported by the library whenever you compute the parameterized correlation matrix.2

VSL_SS_ED_STREAM_QUANT_PARAMS_N

i

Address of a variable that holds the number of parameters of a quantile computation method for streaming data

Required. Set to the number of quantile computation parameters, VSL_SS_SQUANTS_ZW_PARAMS_N.

VSL_SS_ED_STREAM_QUANT_PARAMS

d, s

Address of an array of parameters of a quantile computation method for streaming data

Required. Set the entries of the array according to the description in "Computing Quantiles for Streaming Data" in the Intel® oneAPI Math Kernel Library (oneMKL) Summary Statistics Application Notes document [SS Notes].

VSL_SS_ED_STREAM_QUANT_ORDER_N

i

Address of a variable that holds the number of quantile orders for streaming data

Required. Positive integer value.

VSL_SS_ED_STREAM_QUANT_ORDER

d, s

Address of an array of quantile orders for streaming data

Required. Set entries of the array to values from the interval (0,1). Provide this parameter whenever you compute quantiles.

VSL_SS_ED_STREAM_QUANT_QUANTILES

d, s

Address of an array of quantiles for streaming data

None.

VSL_SS_ED_SUM

d, s

Address of array of sums

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_2R_SUM

d, s

Address of array of raw sums of 2nd order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_3R_SUM

d, s

Address of array of raw sums of 3rd order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_4R_SUM

d, s

Address of array of raw sums of 4th order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_2C_SUM

d, s

Address of array of central sums of 2nd order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_3C_SUM

d, s

Address of array of central sums of 3rd order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_4C_SUM

d, s

Address of array of central sums of 4th order

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_CP

d, s

Address of cross-product matrix

Optional. Set entries of the array to meaningful values (typically zero) if you intend to compute a progressive estimate. Otherwise, do not initialize the array.

VSL_SS_ED_MDAD

d, s

Address of array of median absolute deviations

None.

VSL_SS_ED_MNAD

d, s

Address of array of mean absolute deviations

None.

VSL_SS_ED_SORTED_OBSERV

d, s

Address of the array that stores sorted results

None.

VSL_SS_ED_SORTED_OBSERV_STORAGE

i

Address of a variable that holds the storage format of an output matrix

Required. Provide a supported storage format whenever you specify sorting of the observation matrix.

  1. See Table: "Storage format of matrix of observations and order statistics" for storage formats.

  2. See Table: "Storage formats of a variance-covariance/correlation matrix" for storage formats.