Latvian WordNet is a relational lexico-semantic dictionary, based on Tezaurs.lv lexical data. As of April 2022, Latvian WordNet contains 11446 words , separated into 15790 word senses. The word senses are linked into 8759 synonym sets, and there are 5967 relations between them. The synonym sets are also linked to the English Princeton WordNet. The set of semantic relations in Latvian WordNet include synonymy, antonymy, hyponymy, meronymy, gradation and generic similarity. The words included in the Latvian WordNet are selected according to frequency analysis of the word occurrences in the Balanced Corpus of Modern Latvian.
We are currently enriching Latvian WordNet also with derivational semantic links, creating the first structured resource for the Latvian language. It consists of two types of derivational links: semantic and morphological, providing and multi-layered view of the derivation process. Morphological links list the base and formatives of each derivative, whereas the semantic links show the sense relations between motivation and derivative words. There are currently about 1000 morphological links and 1600 semantic links.
Latvian WordNet dataset can be browsed as part of the Tezaurs.lv online dictionary and also the dataset can be downloaded free of charge from this site. The development of Latvian WordNet is done by Institute of Mathematics and Computer Science.