Wednesday, September 07, 2011

Several basic statistics terms

Deviation: difference between observation and mean of the population, e.g Xi-E(X)

Residual: the difference between observed value and predicted/estimated value.

Error: the difference btw the observed value and the real value (which is usually unobservable)

MSE (Mean square error): the E[(X-X')^2], where X' is the real value of X.

Standard deviation: square root of the mean squared deviation. eg. sqrt(E[(Xi-u)^2]). It's square root of variance (see below)

Variance: a measurement of the dispersion for a set of numbers/observations (e.g. how far they are spread out from each other). If they have a mean, then variance is defined as E[(Xi-u)^2], where u is E(X). It's square of standard deviation. 

No comments:

Post a Comment