Semantic search on Turkish news domain with automatic query expansion

Demir, Tuğba
In this thesis, semantic search on Turkish news domain with query expansion is proposed. Our aim is to provide the user with the most relevant documents related to their entered keywords. Our system uses data sources from Turkish news websites such as Hürriyet, Milliyet, Sabah, etc. Our system extends the user’s query with word embeddings and semantic relatedness. Furthermore, named entities, containing precious information, are extracted from news sources and user query and ranked to return on top of the results. In the rest of search process, it relies on traditional information retrieval (IR) techniques. For Turkish language, to the best of our knowledge, our system is the first attempt to use such search and query extension techniques on news data.