Information extraction and integration for enriching cultural heritage collections

Abstract

Cultural heritage plays an important role in preserving social characteristics and knowledge for future generations. To provide long-term access to these resources, many cultural materials are today archived digitally. The problem arises when each cultural archive, which has own a large database, has been collected with the same cultural types, but different proposes, Therefore, there are various metadata standards that come from each of these archives, making it difficult to enhance, refine, or even improve raw data. This necessitates the need for a novel framework to integrating various subjects and metadata standards as well as extracting relationship among archives for enriching information retrieval. In this paper, we propose a new approach for discovery semantic relations between entities from articles using Wikipedia and various cultural heritage archives as resources. There are (1) dictionary extraction patterns used for extracting terms and meaning for creating a cultural heritage dictionary and (2) semantic relation extraction for extraction relation following question words. For enriching cultural information, the method for enriching cultural heritage information with the result of semantic relation extraction is presented using semantic string similarity matching. An evaluation of different domains shows high performance of the proposed approach.