Pushing on Text Readability Assessment: A Transformer Meets Handcrafted Linguistic Features

Benchmark Model Rank Results
text-classification-on-onestopenglishRoBERTa-RF-T1 hybrid#1Accuracy (5-fold): 0.990
text-classification-on-weebit-readabilityBART-RF-T1 hybrid#2Accuracy (5-fold): 0.905