Intelligent Teaching of Reading and Writing Skills in Higher Education Languages by Integrating Data Mining Technology

In this paper, we first construct a reading comprehension model based on BERT and EDA strategies and, at the same time, propose sliding-window dynamic cropping to reduce the chapter material to a range that BERT can adapt. Secondly, the TF-IDF model for writing material sentences is proposed, and the algorithm is further improved using TFC. A material sentence quality evaluation mechanism is proposed to extract the most useful sentences in writing based on Bi-Attention. Then, the intelligent teaching system is built and analyzed functionally. Finally, a controlled experiment is established to evaluate the intelligent teaching model. According to the study, the experimental class’s reading comprehension scores increased from 23.83 to 33.25, while the control class’s scores did not improve significantly. As for writing, the five aspects of writing improved in the experimental class, and the posttest percentages were 1.0, 0.9, 0.8, 0.8, 0, and 0.98, respectively. Intellectualized teaching can improve both reading comprehension and writing ability.