STATISTICAL PATTERN CLASSIFICATION OF DIRECT SPEECHES IN CHILDREN STORIES

This paper presents a study of three languages, namely Assamese, Bengali and English. The main objective of this study is to pattern classification of direct speeches with special reference to children's stories, in order to find the distinction among all these languages. We consider only the children's stories because, they are found to be similar all over the world with different flavours produced by different cultures, languages and time. We have taken 30 Assamese stories from ‘Burhi Aai'r Xaadhu'(literary translated to Grandma's tales), 27 Bengali stories from “Tuntunir Boi” (Book of the tailor-bird), 62 English stories from Grimm's fairy tales and 16 English stories from Anderson's fairy tales for collecting data. Detailed statistical analyses have been performed by quantifying the direct speeches and presenting them graphically. Non-parametric approaches have been used to test the significant differences among the direct speeches under consideration. It has been shown that there exist significant differences among the direct speeches written by different authors in different languages. Kolmogorov Goodness –of- Fit test, Kruskal-Wallis test and Squared Ranks Test were used for this purpose.