Developer Reference

  • 0.10
  • 10/21/2020
  • Public Content
Contents

Summary Statistics Task Computation Routines

Task computation routines calculate statistical estimates on the data provided and parameters held in the task descriptor. After you create the task and initialize its parameters, you can call the computation routines as many times as necessary. Table
"Summary Statistics Estimates Obtained with
vslSSCompute
Routine"
lists the
respective
statistical estimates.
The Summary Statistics computation routines do not signal floating-point errors, such as overflow or gradual underflow, or operations with
NaNs
(except for the missing values in the observations).
Summary Statistics Estimates Obtained with
vslSSCompute
Routine
Estimate
Support of Observations Available in Blocks
Description
VSL_SS_MEAN
Yes
Computes the array of means.
VSL_SS_SUM
Yes
Computes the array of sums.
VSL_SS_2R_MOM
Yes
Computes the array of the 2
nd
order raw moments.
VSL_SS_2R_SUM
Yes
Computes the array of raw sums of the 2
nd
order.
VSL_SS_3R_MOM
Yes
Computes the array of the 3
rd
order raw moments.
VSL_SS_3R_SUM
Yes
Computes the array of raw sums of the 3
rd
order.
VSL_SS_4R_MOM
Yes
Computes the array of the 4
th
order raw moments.
VSL_SS_4R_SUM
Yes
Computes the array of raw sums of the 4
th
order.
VSL_SS_2C_MOM
Yes
Computes the array of the 2
nd
order central moments.
VSL_SS_2C_SUM
Yes
Computes the array of central sums of the 2
nd
order.
VSL_SS_3C_MOM
Yes
Computes the array of the 3
rd
order central moments.
VSL_SS_3C_SUM
Yes
Computes the array of central sums of the 3
rd
order.
VSL_SS_4C_MOM
Yes
Computes the array of the 4
th
order central moments.
VSL_SS_4C_SUM
Yes
Computes the array of central sums of the 4
th
order.
VSL_SS_KURTOSIS
Yes
Computes the array of kurtosis values.
VSL_SS_SKEWNESS
Yes
Computes the array of skewness values.
VSL_SS_MIN
Yes
Computes the array of minimum values.
VSL_SS_MAX
Yes
Computes the array of maximum values.
VSL_SS_VARIATION
Yes
Computes the array of variation coefficients.
VSL_SS_COV
Yes
Computes a covariance matrix.
VSL_SS_COR
Yes
Computes a correlation matrix. The main diagonal of the correlation matrix holds variances of the random vector components.
VSL_SS_CP
Yes
Computes a cross-product matrix.
VSL_SS_POOLED_COV
No
Computes a pooled covariance matrix.
VSL_SS_POOLED_MEAN
No
Computes an array of pooled means.
VSL_SS_GROUP_COV
No
Computes group covariance matrices.
VSL_SS_GROUP_MEAN
No
Computes group means.
VSL_SS_QUANTS
No
Computes quantiles.
VSL_SS_ORDER_STATS
No
Computes order statistics.
VSL_SS_ROBUST_COV
No
Computes a robust covariance matrix.
VSL_SS_OUTLIERS
No
Detects outliers in the dataset.
VSL_SS_PARTIAL_COV
No
Computes a partial covariance matrix.
VSL_SS_PARTIAL_COR
No
Computes a partial correlation matrix.
VSL_SS_MISSING_VALS
No
Supports missing values in datasets.
VSL_SS_PARAMTR_COR
No
Computes a parameterized correlation matrix.
VSL_SS_STREAM_QUANTS
Yes
Computes quantiles for streaming data.
VSL_SS_MDAD
No
Computes median absolute deviation.
VSL_SS_MNAD
No
Computes mean absolute deviation.
VSL_SS_SORTED_OBSERV
No
Sorts the dataset by the components of the random vector
ξ
.
Table
"Summary Statistics Computation Method"
lists estimate calculation methods supported by
Intel® oneAPI Math Kernel Library
. See the
Intel® oneAPI Math Kernel Library
Summary Statistics Application Notes
document [SS Notes] for a detailed description of the methods.
Summary Statistics Computation Method
Method
Description
VSL_SS_METHOD_FAST
Fast method for calculation of the estimates:
  • raw/central moments
    /sums
    , skewness, kurtosis, variation, variance-covariance/correlation
    /cross-product
    matrix
  • min/max/quantile/order statistics
  • partial variance-covariance
  • median/mean absolute deviation
VSL_SS_METHOD_FAST_USER_MEAN
Fast method for calculation of the estimates given user-defined mean:
  • central moments
    /sums
    of 2-4 order, skewness, kurtosis, variation, variance-covariance/correlation
    /cross-product
    matrix
    , mean absolute deviation
VSL_SS_METHOD_1PASS
One-pass method for calculation of estimates:
  • raw/central moments
    /sums
    , skewness, kurtosis, variation, variance-covariance/correlation
    /cross-product
    matrix
  • pooled/group covariance matrix
VSL_SS_METHOD_TBS
TBS method for robust estimation of covariance and mean
VSL_SS_METHOD_BACON
BACON method for detection of multivariate outliers
VSL_SS_METHOD_MI
Multiple imputation method for support of missing values
VSL_SS_METHOD_SD
Spectral decomposition method for parameterization of a correlation matrix
VSL_SS_METHOD_SQUANTS_ZW
Zhang-Wang (ZW) method for quantile estimation for streaming data
VSL_SS_METHOD_SQUANTS_ZW_FAST
Fast ZW method for quantile estimation for streaming data
VSL_SS_METHOD_RADIX
Radix method for dataset sorting
You can calculate all requested estimates in one call of the routine. For example, to compute a kurtosis and covariance matrix using a fast method, pass a combination of the pre-defined parameters into the
Compute
routine as shown in the example below:
... method = VSL_SS_METHOD_FAST; task_params = VSL_SS_KURTOSIS|VSL_SS_COV; … status = vsldSSCompute( task, task_params, method );
To compute statistical estimates for the next block of observations, you can do one of the following:
  • copy the observations to memory, starting with the address available to the task
  • use one of the appropriate Editors to modify the pointer to the new dataset in the task.
The library does not detect your changes of the dataset and computed statistical estimates. To obtain statistical estimates for a new matrix, change the observations and initialize relevant arrays. You can follow this procedure to compute statistical estimates for observations that come in portions. See Table
"Summary Statistics Estimates Obtained with
vslSSCompute
Routine"
for information on such observations supported by the
Intel® oneAPI Math Kernel Library
Summary Statistics estimators.
To modify parameters of the task using the Task Editors, set the address of the targeted matrix of the observations or change the respective vector component indices. After you complete editing the task parameters, you can compute statistical estimates in the modified environment.
If the task completes successfully, the computation routine returns the zero status code. If an error is detected, the computation routine returns an error code. In particular, an error status code is returned in the following cases:
  • the task pointer is
    NULL
  • memory allocation has failed
  • the calculation has failed for some other reason
You can use the
NULL
task pointer in calls to editor routines. In this case, the routine is terminated and no system crash occurs.

Product and Performance Information

1

Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserverd for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804