What does Skedzy mean?

The word “Skedzy” derives from an apocope of the statistical term skedasticity. Skedasticity describes the variance or dispersion of a set of random values. The term skedasticity is also spelled scedasticity. The spelling “skedasticity” gives homage to the original Greek roots of the word.

The statistician uses the term homoskedastic to describe measured values that have a constant variance. Consider a linear regression: y = a*x + b. With a simple linear regression, the parameters “a” and “b” are estimated by assuming that the variance of the observed values of y is constant with respect to x.

On the other hand, the variance might not be constant. It might increase as x increases. It might fluctuate irregularly. When the variance is not constant, the data are heteroskedastic. In this case, a more complicated statistical model will estimate the parameters a, b and the parameters that describe the variance function.

Why should variance and skedasticity be of interest to supporting IT infrastructure?

Web servers, application servers, database servers, and operating systems often yield a great deal of data that is useful to solving problems and analyzing performance. Oftentimes, the data varies inexplicably because it is shaped by multiple, unmeasured factors. For simplicity, this variation can be modeled as random. The goal: incorporate the skedasticity of the random variations into a sufficiently robust model. This model can inform decisions about resource allocation, or it can spur discussions about measuring other factors. The data from these measurements can support an enhanced model with less variance.

apocope – the word that remains after clipping the end.