Design and Application of an English-assisted Teaching System based on Recurrent Neural Network-Long Short-Term Memory

English education is emerging in a path of adapted learning, and self-study, that is not limited through space and time, is supported by the modern information technology, especially the network technology. The traditional teaching techniques of teachers, classrooms, books are unable to meet requirement of information society. In this research, Deep Learning (DL) technique named as Recurrent Neural Network-Long Short-Term Memory (RNN-LSTM) is proposed for English assisted teaching system. At the language practice stage, the help system act as learning, talking in English with learners. The secondary system is organized to reproduce learner English stylishness which plays a main role in task situation. The next step of the linguistic practice part involves the implementation of multiple roles for the system that match the user’s definition of identity. It does import discussion topics covering different domains and is designed for the modification and restructuring of such ideas. The performance of RNN-LSTM is calculated by metrics like f1-score, precision, accuracy, and recall. The RNNLSTM achieves 98.64% f1-score, 98.83% precision, 99.27% accuracy, 98.39% recall when compared to Machine Learning-based Gaussian Process Model (MLGPM).