home edit page issue tracker

This page pertains to UD version 2.

Treebank Statistics: UD_Russian-Taiga: Features: NumForm

This feature is language-specific. It occurs with 5 different values: Combi, Cyril, Digit, Roman, Word.

22560 tokens (1%) have a non-empty value of NumForm. 1740 types (1%) occur at least once with a non-empty value of NumForm. 1305 lemmas (2%) occur at least once with a non-empty value of NumForm. The feature is used with 2 part-of-speech tags: NUM (12849; 1% instances), ADJ (9711; 1% instances).

NUM

12849 NUM tokens (100% of all NUM tokens) have a non-empty value of NumForm.

The most frequent other feature values with which NUM and NumForm co-occurred: NumType=Card (12069; 94%), Number=EMPTY (10948; 85%), Gender=EMPTY (8705; 68%).

NUM tokens may have the following values of NumForm:

Paradigm 2CombiDigit
Case=Gen|Gender=Masc|NumType=Card2-х
Case=Gen|NumType=Card2-х, 2х
NumType=Card2
NumType=Sets2

NumForm seems to be lexical feature of NUM. 98% lemmas (531) occur only with one value of NumForm.

ADJ

9711 ADJ tokens (6% of all ADJ tokens) have a non-empty value of NumForm.

The most frequent other feature values with which ADJ and NumForm co-occurred: Degree=EMPTY (9711; 100%), Gender=EMPTY (6709; 69%), Case=EMPTY (5839; 60%), Number=EMPTY (5839; 60%).

ADJ tokens may have the following values of NumForm:

Paradigm IRomanWord
Case=Nom|Gender=Masc|Number=SingI
II

NumForm seems to be lexical feature of ADJ. 99% lemmas (838) occur only with one value of NumForm.

Relations with Agreement in NumForm

The 10 most frequent relations where parent and child node agree in NumForm: ADJ –[nmod]–> ADJ (805; 92%), NUM –[nmod]–> NUM (405; 100%), NUM –[conj]–> NUM (302; 100%), ADJ –[conj]–> ADJ (162; 72%), NUM –[compound]–> NUM (133; 100%), ADJ –[compound]–> NUM (36; 62%), NUM –[list]–> NUM (21; 84%), NUM –[parataxis]–> NUM (16; 84%), NUM –[nummod:gov]–> NUM (13; 100%), ADJ –[nmod]–> NUM (6; 86%).