Abstract

The objective of this paper is the assessment of the comprehensibility in museum panels, considering that they are addressed to all audiences and written in standard language. The indicators proposed in the scientific bibliography on comprehensibility are compiled. Subsequently, their application to the museum panels is analyzed. A corpus is built with the texts of panels from six museums and the difficulty of the panels is evaluated by means of classical metrics and tests on users’ perception. The ability of these classic metrics to predict the difficulty for the user is computed with automatic learning methods. Linguistic and term familiarity indicators have been added to improve the accuracy of the assessment. The most effective way to predict the degree of comprehensibility is a hybrid model of classical, linguistic and term familiarity indicators.