Intel® Summary Statistics Library Survey (Intel® Math Kernel Library)

Intel® Summary Statistics Library includes implementations of basic algorithms used to summarize a set of observations in order to communicate the largest amount as simply as possible, all highly optimized for performance on the Intel architecture.
Available at http://whatif.intel.com

We’d like your input on potential new functionality in future versions of the Intel® Math Kernel Library. This survey is designed to take no more than 15 minutes. Thank you for taking the time to help us make products that help you to succeed.

The Intel Math Kernel Library Team

Which of the following categories best describes your application area(s)?

	Energy
	Manufacturing/Automotive
	Finance
	Life Sciences
	Digital Media and Gaming
	Healthcare
	Communications
	Academic; No vertical above
	Other (please specify)

What functionality in Intel® Summary Statistics Library is useful to you?

	Basic estimates (algebraic/central moments/kurtosis/skewness/variation)
	Quantiles, order statistics, quantiles for streaming data, minimum, maximum
	Algorithms for variance-covariance/correlation matrix
	Detection of outliers in datasets
	Support of missing values in datasets
	Parameterization of correlation matrix
	Other (please specify)

Which version of the library did you evaluate?

	Version 1.0.0
	Version 1.0.1

Do you want to see the capabilities you used in Intel® Summary Statistics Library appear in future versions of Intel® Math Kernel Library?

	Yes, definitely
	Yes, I’d find it desirable
	No, I don’t need it
	No, I’m using some other library
	Don’t care

Are algorithms of Intel® Summary Statistics Library relevant to your applications?

	Yes
	No
	Need to investigate (please explain)

Do you currently use any statistical packages in your applications?

	IMSL*
	NAG*
	R*
	SAS*
	Proprietary code
	Other (Please specify)

What are your impressions on the API in Intel® Summary Statistics Library?

Easy to use	Easily integrates into applications	Not intuitive	Relatively hard to understand	Unusable	Please comment on your experience:

I prefer that the SSL API need to be most similar to of the following packages?

	IMSL*
	NAG*
	R*
	SAS*
	Other (please specify)

From which languages did you call and/or desire to call Intel® Summary Statistics Library? (In the first column, indicate the % of your current usage of each language. In the 2nd column, indicate the % of your desired future (1-3 years) usage of each language. Each column should total 100%.)

	Current Usage	Future Usage (1-3 years)
C# (.NET)
F# (.NET)
Java
C
C++
FORTRAN 95
Others (please specify)

Please describe your impressions on the documentation made available with the Intel® Summary Statistics Library.

	Very easy	Easy	Average usage	Difficult	Very difficult	Suggestions to improve
Is it easy to get started with the library?
Is it easy to find necessary information in the documentation?
Are the examples demonstrative?

For your applications, which of the following algorithms in Intel® Summary Statistics Library need emphasis on performance?

	Basic estimates (algebraic/central moments/kurtosis/skewness/variation)
	Quantiles, order statistics, quantiles for streaming data, minimum, maximum
	Algorithms for variance-covariance/correlation matrix
	Detection of outliers in datasets
	Support of missing values in datasets
	Parameterization of correlation matrix
	Don’t care about performance

Which algorithms do you think are missing in the Intel® Summary Statistics Library that may be useful in your applications?

If we can contact you to get more details on your input please fill in your name and email below. Note that this is completely optional.

Name (optional)
Email (optional)