Title :
A multisample criterion for changepoint analysis of texts
Author :
Zakrevskaya, N.S.
Author_Institution :
Novosibirsk State Tech. Univ., Russia
fDate :
26 June-2 July 2005
Abstract :
We construct a criterion to differ homogeneous and non-homogeneous texts. This criterion is based on triplets´ frequencies analysis: we find the most deviated corresponding empirical bridge and analyze its deviation. The approach can differ homogeneous and non-homogeneous texts.
Keywords :
natural languages; text analysis; homogeneous texts; nonhomogeneous texts; text changepoint analysis; text identification; triplet frequencies analysis; Bridges; Frequency conversion; Libraries; Sections; Testing;
Conference_Titel :
Science and Technology, 2005. KORUS 2005. Proceedings. The 9th Russian-Korean International Symposium on
Print_ISBN :
0-7803-8943-3
DOI :
10.1109/KORUS.2005.1507893
