Regression model of rank –frequency data of Tamil text

Abstract

The application of Zipf’s law is universal not only in linguistics but also in various other areas. Mandelbrot modified Zipf law as Zipf Mandelbrot law and it is further we proposed a modification of the ZM law for modeling rank frequency- data of linguistic text. Our model generalized ZM law into a linear regression model involving arbitrary order of Zipfian rank of words in a text .The performance of the proposed model is studied for an English text and it shown to compare favorably with that of Z-M law using Chi-Square goodness of fit test. In this paper we have applied to Tamil text and its performance is also up to the mark and it is been proved by the Chi-Square test and it addresses mainly the lower ranks, we propose to extend the work to higher order ranks using LNRE model in the future.