Exposing Predictive Analytics through Natural Language

Data processing applications often have non-experts as audience, which has led to an increasing interest into verbalizing data into natural language. However, verbalizations often focus on observed or predicted facts, while predictive technologies in addition generate information such as probabilities and confidence scores. This gives rise to data that goes beyond facts alone and that also has to be communicated to its users in order to enable a correct interpretation of the predictions. In this paper we explore the natural language needed when applications offer predictive analytics technology to their users, and demonstrate how to modularly implement the grammars needed to verbalize common aspects of predictions.