Assessing English Language Writing and Readability Skills using Long Short-Term Memory Model

In English writing courses, students’ ability to write effectively relies on how well they comprehend the texts that are used in teaching. In order to determine the comprehension level, it is fundamental to have a readability score measurement of written texts. The readability score refers to a numeric measure that defines the level at which a learner can comprehend a text. Traditional readability formulas such as Flesch Reading Ease, Dale-Chall and FleschKincaid Grade Level and other formulas are widely used to assess the text’s readability score. In contrast, the Long ShortTerm Memory is a variant of Deep Learning Recurrent Neural Network that is designed to handle sequences of data such as text and time series. Long ShortTerm Memory can extract several features from the text including sentence patterns and readability scores. This research proposes a Long Short-Term Memory model that can predict the readability score of texts that is scored using traditional formulas.