Article: Evaluation of teaching quality in database courses based on domain-adaptive transfer learning Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.24 pp.49 - 64 Abstract: The distribution of teaching data varies among database courses, and traditional methods are often difficult to deal with such domain differences effectively. For this reason, this paper firstly utilises BERT model for embedding learning of teaching feedback text, and then extracts local and global features of the text through convolutional neural network (CNN) and long short-term memory (LSTM) network respectively, and enhances the text features through the attention mechanism. On this basis, the domain adaptive transfer learning algorithm is adopted to achieve the characteristic distribution migration alignment of the text source topic and objective topic, and minimise the scoring difference between different classifiers through consistency constraints, so as to assess the teaching quality more accurately. Simulation results show that the classification accuracy of the offered method is 94.39%, which demonstrates a substantial enhancement over the benchmark method. Inderscience Publishers - linking academia, business and industry through research

Title: Evaluation of teaching quality in database courses based on domain-adaptive transfer learning

Authors: Xuesong Yang

Addresses: College of Electrical Engineering, Northwest Minzu University, Gansu 730000, China

Abstract: The distribution of teaching data varies among database courses, and traditional methods are often difficult to deal with such domain differences effectively. For this reason, this paper firstly utilises BERT model for embedding learning of teaching feedback text, and then extracts local and global features of the text through convolutional neural network (CNN) and long short-term memory (LSTM) network respectively, and enhances the text features through the attention mechanism. On this basis, the domain adaptive transfer learning algorithm is adopted to achieve the characteristic distribution migration alignment of the text source topic and objective topic, and minimise the scoring difference between different classifiers through consistency constraints, so as to assess the teaching quality more accurately. Simulation results show that the classification accuracy of the offered method is 94.39%, which demonstrates a substantial enhancement over the benchmark method.

Keywords: database curriculum; teaching quality evaluation; BERT model; domain adaptation; transfer learning.

DOI: 10.1504/IJICT.2025.147132

International Journal of Information and Communication Technology, 2025 Vol.26 No.24, pp.49 - 64

Received: 06 May 2025
Accepted: 24 May 2025
Published online: 10 Jul 2025 *

Title: Evaluation of teaching quality in database courses based on domain-adaptive transfer learning

Keep up-to-date