Development of a Network Service for Assessing the Complexity of Natural Language Texts

An important step in learning foreign languages is writing an essay. The assessment of their complexity can be considered an indirect indicator of the level of language proficiency. However, such an assessment is a time-consuming process that requires a lot of time and effort from experts. In turn, students need a tool to evaluate the work done before passing it. Currently existing services mostly use simple readability metrics and do not provide a comprehensive assessment of the text. The purpose of this work is to develop a service that evaluates the complexity of texts in English. The study analyzed the main currently existing statistical methods of machine learning used for classification. Based on these methods, models are trained to evaluate the complexity of texts from the REALEC corpus. A service has been developed that uses the best of the received models and gives feedback on improving the evaluated text. The results of this work can be used by students studying English and their teachers.