Design of English teaching quality assessment system model based on forward rule fast matching algorithm

Abstract

Education is vital for talent development, and assessing English teaching quality helps teachers improve their methods. However, traditional assessment models are inefficient and time-consuming. To address this, this study proposes a new teaching quality assessment model using a forward rule fast-matching algorithm. This approach improves data processing efficiency by removing irrelevant data through a rule-driven data cleaning. The model integrates a deep belief network with the fast-matching algorithm to create an effective evaluation tool, along with a student performance prediction model. Finally, for model validation, the study used a self-collected dataset that included English performance data from 286 students, which was later expanded to include English performance data from 3500 students from 10 universities of different sizes. Experimental results show that the model performs well, achieving a PR curve area of 0.87 and an F1 value of 0.93. After regularization, the model reduced time consumption by 6.18% and space usage by 3.23%. The average matching time was 319 ms, with 927 matches on average—both better than existing methods. Moreover, the difference between actual and predicted English scores was small, with detection accuracy above 93%. In conclusion, the proposed model offers higher efficiency, accuracy, and reliability in evaluating English teaching quality. The research contributes by reducing time costs and improving evaluation effectiveness, offering a practical solution for real-world teaching assessments. It also advances deep learning by combining deep belief networks with forward rule fast-matching techniques.