Statistical analysis of sentences with phraseological structures: association measure of mutual information
Hanna Sytar · Ukrainian Linguistics · 2016
The article is devoted to the statistical analysis of the sentences with phraseological structures of the Ukrainian language. Expediency of application of statistical criterion of mutual information was substantiated for determining the nonrandom measure of a certain sequence of words in the text. There were provided the results of mutual information computation for the models of sentences with phraseological structures according to the data of Ukrainian National Linguistic Corpus. There was proved that all the analyzed sentences models have a high degree of components non-randomness that make up invariable part of the sentence. There were suggested the computation of the modified measure of МІ – МІ3 for the researched units. Obtained data was compared with the corresponding measures of MI and МІ3 for lexical phraseologisms and non-idiomatic sentences. There were singled out the factors that affect the correctness of calculations performed.