Linguistic Features for Readability Assessment

Abstract

Readability assessment aims to automatically classify text by the level appropriate for learning readers. Traditional approaches to this task utilize a large variety of linguistically motivated features paired with simple machine learning models. More recent methods have improved performance by discarding these features and utilizing deep learning models. This thesis attempts to combine these two approaches with the goal of improving overall model performance. My primary method involves incorporating the output of a deep learning model as a feature itself, used in conjunction with linguistic features. Evaluating on two large readability corpora, I find that this fused approach is ineffective, failing to improve upon state-of-the-art performance. These results suggest that other avenues of research would be more fruitful in improving readability assessment.