Type of statistics

Mtas can produce several type of statistics, e.g. for positions, tokens or spans.

In general, statistics of type basic will require less resources than statistics of type advanced, whereas statistics of type advanced will require less than these of type full. If multiple statistics are possible and required, the parameters can be seperated with a comma.

Parameter Description Type
n number of documents basic
sum total basic
mean mean basic
min minimum advanced
max maximum advanced
sumsq sum of squares advanced
sumoflogs sum of logs advanced
geometricmean geometric mean advanced
quadraticmean quadratic mean advanced
variance variance advanced
populationvariance population variance advanced
standarddeviation standard deviation advanced
skewness skewness full
median median full
kurtosis kurtosis full
all all of the above full



When not obligatory, usually "n,sum,mean" will be the default.


Besides these parameters, a distribution can requested with (optional) arguments, optionally combined with one or more of the parameters listed above.

Parameter Type Arguments Description Obligatory
distribution() full start start value no
end end value no
step size of intervals no
number number of intervals no