Research on the Teaching Reform of “Communication and Writing” in Colleges and Universities under the Background of Melting Media

Abstract

Students frequently encounter challenges in vocabulary selection, clarity of expression, and the construction of lengthy, under-informed essays during communication and writing tasks. To address these issues, this paper introduces a novel approach utilizing sentence recommendation to enhance students’ communicative and writing competencies. Specifically, the proposed model employs a Probabilistic Latent Semantic Analysis (PLSA) topic model to calculate the similarity of Chinese sentences and recommends contextually relevant sentences to students, thereby guiding their communication and writing efforts. Additionally, the Latent Dirichlet Allocation (LDA) topic model is applied to analyze and mine the corpus at both word and sentence levels, culminating in the creation of comprehensive and varied practice documents through thoughtful integration. Building on the capabilities of these models, we devise a pedagogical strategy for the course “Communication and Writing”. Empirical testing reveals significant improvements; the average score of the experimental class’s pre-test exceeded the post-test by 4.01 points in terms of self-efficacy in communication and writing tasks. Moreover, the total scores reflecting students’ self-efficacy in communication and writing assignments and skills showed marked enhancement, bolstering their self-efficacy in these areas. Regarding behavioral changes, the proportion of students frequently initiating communication and writing activities surged from 4.5% to 23%, an increase of 18.5%. Ultimately, the instructional strategies outlined in this study provided students with effective practice tools, significantly elevating their cognitive and sentence quality in communication and writing tasks.

Keywords

• communication and writing
• plsa
• sentence recommendation
• lda topic modeling
• 35b27