On Digitalizing the Geographical Dictionary of Polish Kingdom Published in 1880

Abstract

Printed encyclopedic texts provide us with great quality and organized textual data but make us face several problems. Most of the modern data extractors are indexers that create rather simple databases. Such type of database provides keyword search and simple links between subjects (based on co-occurrences). The most advanced extractors are supported by predefined ontologies which help to build relations between concepts. In this paper we study the problem of digitalizing the Geographical Dictionary of Polish Kingdom published in 1880. We address two kinds of problems: technical problems of converting scanned pages of that dictionary to reliable textual data and theoretical challenges of organizing that data into knowledge. Our solution for organizing data into knowledge is based on defining an appropriate ontology and rules for converting textual data to ontological knowledge. We describe methods of extracting simple information like keywords and bootstrapping it into higher level relations and discuss their possible uses.