Automatic Creation of Knowledge Graphs from Digital Musical Document Libraries
Most of the current musicological knowledge is present in text documents. Greats efforts have been done in order to digitize and make them available. However, digital documents are mainly stored as raw text, and implicit knowledge is not understandable by computers. We propose a method to automatically extract knowledge from Digital Document Libraries, by using Information Extraction techniques. The extracted knowledge is represented as a knowledge graph and visualized in an interactive tool. The method is tested on artist biographies from Grove Music Online.
(2014). Automatic Creation of Knowledge Graphs from Digital Musical Document Libraries. Conference in Interdisciplinary Musicology.