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 |
Examples
n,sum mean,standarddeviation
When not obligatory, usually "n,sum,mean" will be the default.
Distribution
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 |
Examples
distribution() distribution(start=0,end=1,number=10) mean,standarddeviation,median,distribution(start=0,end=1,number=10)