Institutional Repository of Vilnius University
Semantinės paieškos ir atvirkštinio indeksavimo taikymas lietuvių kalbos žodynuose /
Abstract
dc:descriptionThis work analyzes semantic search, implemented using semantic web and search with inverted index. Results scoring function, semantic web and ways to implement semantic search are reviewed. A study of search of General Lithuanian Language Dictionary is conducted. Software is created which does search and semantic search and indexes data. For search with inverted index, Solr search engine is used and for semantic search SPARQL queries and Lucene search engine is used. The aim of study was to compare different scoring functions, different text analyzers, different indexing strategies, search vs semantic search results precision and/or search speed. Study showed that different scoring functions: BM25Similarity, ClassicSimilarity and DFRSimilarity didn’t have much of an impact on result precision and search speed. LithuanianAnalyzer text analyzer had lower precision than standard Solr and Lucene analyzers. Semantic search results precision and search speed was dependent on the data indexing strategies, where fastest and most precise was indexing strategy where 1 document is 1 triplet, least precise results were with indexing strategy where 1 document is 1 entity and slowest search was when it was searched without inverted index. It is useful to use semantic search on General Lithuanian Language dictionary with short queries, because semantic search had higher precision than search with inverted index, but with long queries semantic search speed was very slow compared to search with inverted index.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rudis, Laurynas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD146242312&prefLang=en_US