Chapter 3
Unification of measures
Article outline
- 3.1Introduction: The Kullback-Leibler divergence
- 3.2Characteristics of the KLD
- 3.2.1Directionality
- 3.2.2Independence of sample size
- 3.2.3Normalization
- 3.2.4Smoothing
- 3.3Background and interpretation
- 3.4Application to dispersion
- 3.4.1A first simple example
- 3.4.2A more realistic example
- 3.4.3A (very brief!) second reaction to Nelson (2023)
- 3.5Application to association
- 3.5.1A first simple example
- 3.5.1.1The direction row-to-column
- 3.5.1.2The direction column-to-row
- 3.5.1.3Quick discussion
- 3.5.2A second simple example
- 3.5.2.1The direction row/verb→column/construction
- 3.5.2.2The direction column/construction→row/verb
- 3.5.3A more realistic example
- 3.5.3.1The direction row/verb→column/ditransitive
- 3.5.3.2The direction column/ditransitive→row/verb
- 3.6Application to key words
- 3.6.1A first simple example
- 3.6.1.1The direction row/word→column/corpus
- 3.6.1.2The direction column/corpus→row/word
- 3.6.2A more realistic example
- 3.6.2.1The direction row/word→column/corpus
- 3.6.2.2The direction column/corpus→row/word
- 3.7A brief excursus on contributions to KLD
- 3.7.1Contributions to KLD for keyness
- 3.7.2Contributions to KLD for association
- 3.8Application to concordancing
- 3.8.1A first simple example
- 3.8.2A more realistic example
- 3.8.3Implications for frequency
- 3.9Interesting extensions
- 3.9.1Psycholinguistic applications (incl. prototypicality)
- 3.9.2Synchronic and diachronic corpus homogeneity/comparisons
- 3.10Interim summary
-
Notes