Semantic network
Definition
Semantic network indices capture lexical relationships such as polysemy (multiple meanings per word) and hypernymy (hierarchical relationships between general and specific terms, e.g., flower vs. rose).
Methodology
TAALES calculates semantic network indices using path-based measures in WordNet. Scores reflect either the number of senses a word has (polysemy) or the distance between concepts in a semantic hierarchy (hypernymy). Indices are averaged across parts of speech and path combinations.
Corpus Used
- WordNet (Fellbaum, 1998)
Calculated indices
Polysemy
- Indices:
- content_poly
- poly_noun
- poly_verb
- poly_adj
- poly_adv
Hypernymy – Nouns
- Indices:
- hyper_noun_S1_P1
- hyper_noun_Sav_P1
- hyper_noun_Sav_Pav
Hypernymy – Verbs
- Indices:
- hyper_verb_S1_P1
- hyper_verb_Sav_P1
- hyper_verb_Sav_Pav
Hypernymy – Combined (Nouns & Verbs)
- Indices:
- hyper_verb_noun_s1_p1
- hyper_verb_noun_Sav_P1
- hyper_verb_noun_Sav_Pav
Notes
- S1: Uses the first (most frequent) sense of the word.
- Sav: Averages across all senses of the word.
- P1: Uses the first (most direct) path in the hypernym hierarchy.
- Pav: Averages across all possible paths for the word.
References:
- Fellbaum, C. (Ed.). (1998). WordNet: An electronic lexical database. MIT press.