Notes on the Negative Binomial distribution for word occurrences
Edoardo M. Airoldi
Abstract
Edoardo M. Airoldi
Abstract
rms of number of words; and #, which is the length of a document expressed as a pure number, multiple of the word-length of the reference text, e.g., # = 1.67 for a text 1670 word long if the reference text is a thousand words, i.e., # = 1000. This allows us to express the rate # as the rate of occurrence of a word in a text of length equal to that of the reference text, , conditionally on the desired, or observed, length of the text, #, expressed as a multiple of the word-length of the reference text. References
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
rms of number of words; and #, which is the length of a document expressed as a pure number, multiple of the word-length of the reference text, e.g., # = 1.67 for a text 1670 word long if the reference text is a thousand words, i.e., # = 1000. This allows us to express the rate # as the rate of occurrence of a word in a text of length equal to that of the reference text, , conditionally on the desired, or observed, length of the text, #, expressed as a multiple of the word-length of the reference text. References
Key concepts: Negative binomial distribution, Poisson distribution, Negative multinomial distribution, Mathematics, Poisson binomial distribution, Count data, Binomial distribution, Statistics