Diagnosing Causes of Reading Difficulty using Bayesian Networks

There is a need of matching text difficulty to the expected reading skill of the audience. Readability measures were developed with this objective in mind, first by psycholinguists, and more recently, by practitioners of natural language processing. A common strategy was to extract linguistic features that are good predictors of readability, and then train discriminative classification or regression models that correlate well with human judgment. But correlation does not imply causality, which is a necessary property to explain why documents are not readable. Our objective is to provide mechanisms for text producers to adjust the readability of their content. We propose the use of generative models to diagnose causes of reading difficulty, and bring closer the realization of automatic readability optimization.