This package implements the increasing window entropy estimator (i.e., LZ78) as described in [1], [2] and [3] to estimate the entropy rate of a string of symbols. In the example given here, symbols are words, and the string is a pre-processed text of the German version of the European Parliament Corpus (EPC) [4]. The package provides two main functions: stabilize.estimate() and get.estimate(). The first function calculates the entropy rate for a pre-specified set of token counts. This helps to establish the number of tokens necessary for entropy rates to stabilize. The second function gives a single estimate for a pre-specified number of tokens.


Install either directly from GitHub


or from the tarball

install.packages("Hrate.tar.gz", repos=NULL, type="source")



## load the provided demo corpus: 50 K tokens of the German Europarl 
## punctuation and numbers are removed and word tokens are set to lower case

[1] "wiederaufnahme"  "der"             "sitzungsperiode" "ich"             "erkläre"
[6] "die"             "am"              "freitag"         "dem"             "dezember"

## get the stabilization points via stabilize.estimate()
## "step.size" is an argument specifying step sizes (in number of tokens) at which entropy rates are calculated. "max.length" specifies the maximum number of tokens to be included. "every.word=1" specifies that each word token should be used for estimation. To speed up processing only every 2nd, 3rd, xth word token could be used. Hence, every.word can be assigned any integer between 1 and the step size. "rate" gives the downsampling rate to get SDs over a given number of entropy rate estimations (see Section 4.2 in [3]). converge.estimate returns a S4 object. 
stabilization.rate <- stabilize.estimate(text = deuparl, step.size = 1000, max.length = 50000, every.word = 10, method="downsample",rate = 5)

## output the convergence rate and the associated SDs

## we also expose a get.estimate function to get a single estimate on a text. "every.word""
## and "max.length" are the same arguments as for converge.estimate().
estimate <- get.estimate(text = deuparl, every.word = 10, max.length = 50000)

## plot the stabilization results


  1. Gao, Y.; Kontoyiannis, I.; Bienenstock, E. Estimating the entropy of binary time series: Methodology, some theory and a simulation study. Entropy 2008, 10, 71–99.

  2. Bentz, C.; Alikaniotis D. The word entropy of natural languages, arXiv 2016.

  3. Bentz, C.; Alikaniotis D.; Cysouw, M.; Ferrer-i-Cancho, R. The entropy of words - learnability and expressivity across more than 1000 languages. Entropy 2017.

  4. Koehn, P. Europarl: A parallel corpus for statistical machine translation. In Proceedings of the tenth Machine Translation Summit, Phuket, Thailand, 12-16 September 2005; Volume 5, pp. 79–86.

Travis-CI Build Status

dimalik/Hrate documentation built on May 24, 2019, 4:01 a.m.