On variation of word frequencies in Russian literary texts

摘要: Abstract We study the variation of word frequencies in Russian literary texts. Our findings indicate that standard deviation a word’s frequency across texts depends on its average according to power law with exponent 1 2 α , which shows rarer words have relatively larger degree volatility (that is, higher “burstiness”). A latent factor model has been estimated investigate structure distribution. The suggest dependence can be explained by asymmetry distribution factors.