Predicting the fluency of text with shallow structural features: case studies of machine translation and human-written text

Sentence fluency is an important component of overall text readability but few studies in natural language processing have sought to understand the factors that define it. We report the results of an initial study into the predictive power of surface syntactic statistics for the task; we use fluency assessments done for the purpose of evaluating machine translation. We find that these features are weakly but significantly correlated with fluency. Machine and human translations can be distinguished with accuracy over 80%. The performance of pairwise comparison of fluency is also very high---over 90% for a multi-layer perceptron classifier. We also test the hypothesis that the learned models capture general fluency properties applicable to human-written text. The results do not support this hypothesis: prediction accuracy on the new data is only 57%. This finding suggests that developing a dedicated, task-independent corpus of fluency judgments will be beneficial for further investigations of the problem.