Semantic analysis of the user queries in the Croatian Historical Newspapers Portal log file

Semantic analysis of the user queries in the Croatian Historical Newspapers Portal log file

A historical newspapers digital library contains the basic elements and relations of a digital library, but also some special features and functions necessary for processing its content - huge amount of digitised pages with complex granularity and large amount of (historical) text. Descriptive metadata of newspapers are usually available on the title level only and the full-text search depends on the accuracy of OCR as well as orthographic and semantic issues of historical text. The characteristics of the historical newspapers digital library content pose a challenge for information retrieval and give rise to the following questions: How do users search digitised historical newspapers?, How do they formulate their search queries?, What topics are they looking for?, How do they deal with historical text issues?. Data stored in the historical newspapers digital library search logs can provide some of the answers and help to improve information access. The paper reports on the results of the semantic analysis of the Croatian Historical Newspapers Portal user queries.