Predict the Complexity of Literary Passages Based on Deep Learning

Graded reading is becoming increasingly important in everyday life. For example, recommending books suitable for the reading difficulty of the current age to children can bring great convenience to parents. At the same time, it can also increase children’s interest in reading and prevent them from being stressed because of the difficulty of reading. However, using humans to judge the reading difficulty of texts and then grading them requires more manpower and material resources. We need to use deep learning methods to automatically judge the reading difficulty of text. Our paper proposes a Roberta-based model to rate the difficulty of a given text. First, the data set is preprocessed, and then use the training data set and the validation data set to train and optimize the model. In addition, our model is compared with LSTM and RNN. Experiments show that the model proposed in this paper has better performance on the task of text reading difficulty classification.