I once read that, for example, there are statistical techniques that are applied in social sciences that also are applied for Computer Science and other fields, for example: sampling, hypothesis testing, and so on. I was wondering if there are some statistical specific techniques that are used in the field of Computer Science and related to topics like information systems, software engineering and so on.

I know that there is the field of Computational Statistics, but that is more related to the development of algorithms for statistics, so it is somewhat different.

closed as off-topic by Federico Poloni, Wrzlprmft, Enthusiastic Engineer, Fomite, scaaahu Sep 20 '15 at 2:29

  • This question does not appear to be about academia within the scope defined in the help center.
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 4
    I'm voting to close this question as off-topic because it belongs either on stats.stackexchange.com or cs.stackexchange.com. Either site has a more suitable readership to answer this question than academia.se. – Federico Poloni Sep 19 '15 at 19:03

Although statistical techniques are fundamentally the same, specific approaches to the use of statistics in experimental analysis of algorithms or more broadly in the analysis of computer systems performance (including aspects of hardware, networks, etc.) are an important topic. A couple of references to look at include:

C. C. McGeoch. A Guide To Experimental Algorithmics. Cambridge University Press, 2012.

R. Jain. The Art of Computer Systems Performance Analysis: Techniques for Experimental Design, Measurement, Simulation, and Modeling. Wiley, 1991.


There is nothing special about statistical methods used in computer science compared to statistical methods in other fields. Once a technique has been developed, it can move freely from field to field as people find use for it.

There are techniques that happened to be developed first in computer science, due to the specific applications that drove their development, just as there are specific techniques that happened to be developed first in other fields. Places that you are likely to encounter such techniques include generally "machine learning" and "big data," but again, there is nothing particularly special about those fields, it's just that they happen to be dealing with large masses of data and data-centric problems that tend to drive development of new techniques.


Actually, statistics employed in computer science can be used in nearly any real life problems. This is so since computers were created at first to address them.

But there is a statistical measure that is known as a good practice when measuring accuracy of a proposed model. It's called the F score or F1 score.

F score = 2PR/(P+R)


Precision, P = #true positives / #predicted positives

Recall, R = #true positives / #actual positives

The F (or F1) score is used in places like machine learning where prediction of an event is required. Predicted positives are those in which the event is predicted to occur. True positives relate to the number of instances predicted correctly. Actual positives are the total instances of the concerned event. This event may be an event of cancer in healthcare, defect manufacture, success in gene splicing, etc.

  • Not to be confused with what is usually called an F-statistic. – Michael Hardy Sep 22 '15 at 3:42

Not the answer you're looking for? Browse other questions tagged or ask your own question.