Visual Topical Analysis of Museum Collections

Museums are highly specialized cultural institutions. Obstacles exist between the knowledge and terminology of the museum professionals and that of the general public. Topical analysis of museum collections can reveal topical similarities and differences among museums and facilitate museum tours with recommended professional guides. In this study, 7177 French artworks collected by 90 art museums worldwide were investigated. The Self-Organizing Map SOM technique, an unsupervised artificial neural network method, was applied to visually analyze similarities and differences among the museums. The Treemap technique was also employed on a large dataset to reveal the distribution of the specific themes among the investigated museums. Finally, a comprehensive museum tour recommendation mechanism is established for tourists.