Using Morphological Data in Language Modeling for Serbian Large Vocabulary Speech Recognition

Table 5

Some of the most frequent words, with and without morphology-based suffixes.

Without POS data	With POS data	Explanation
je	je_gl	glagol = verb
i	i_vez	veznik = conjunction
u	da_vez	—
da	u_pred_dat	predlog = preposition
se	se_zam	zamenica = pronoun
na	na_pred_dat	dative/locative
koji	koji_zam_nom_jd_mr	nominative
bi	bi_gl_jd	jednina = singular
Srbije	Srbije_im_gen_jd_mr	imenica = noun, genitive, muški rod = masculine