1　Introduction

This section introduces the definition, characteristics, and current research status of prepositional frame structures, reviews domestic scholars' research findings on their syntactic functions, typological features, and semantic integrity, examines empirical studies in the field of second language acquisition regarding acquisition difficulty, error types, and native language transfer among learners with different native language backgrounds, points out that current research mainly focuses on elementary and intermediate stages as well as European, American, Japanese, and Korean learners, with insufficient attention paid to Southeast Asian advanced learners, and finally clarifies that this study takes advanced learners from Thailand, Vietnam, and Indonesia as subjects, analyzes their acquisition advantages and difficulties based on the HSK Dynamic Composition Corpus, and provides references for teaching.

2　Research Methods

This section introduces the research design for investigating the use of prepositional frame structures by Southeast Asian advanced Chinese learners based on the HSK Chinese Proficiency Test composition corpus. The research addresses five core questions, including the current status of errors, level differences, native language background influence, successful strategy summarization, and teaching method design. The corpus was selected from Beijing Language and Culture University's HSK Dynamic Composition Corpus, limited to international students with Vietnamese, Thai, and Indonesian as their native languages. Learners were divided into elementary, intermediate, and advanced groups based on HSK levels and the "Chinese Proficiency Grading Standards for International Chinese Language Education." Data processing adopted a four-category method of omission, redundancy, misordering, and semantic errors, using Python for word segmentation annotation and error classification, combined with manual verification to obtain 229 valid samples, and then using SPSS for analysis of variance and chi-square tests, with Excel calculating accuracy rates and error ratios to ensure the validity and accuracy of the analysis.

3　Data Analysis

This section systematically examines the accuracy rates, error types, usage frequency, and diversity of prepositional frame structure acquisition among Southeast Asian Chinese learners at different proficiency levels. The study found that accuracy rates improve with increasing language proficiency, with elementary (54.79%) and intermediate (76.70%) levels significantly lower than advanced (92.45%); omission and semantic errors account for higher proportions at elementary and intermediate levels, while redundancy errors persist stubbornly at the advanced stage. There is a significant correlation between nationality and error type distribution: Thai learners show high proportions of omission, redundancy, and semantic errors; Vietnamese learners prominently display omission, misordering, and semantic errors; Indonesian learners concentrate on omission and semantic errors. Two-way ANOVA shows that language level significantly affects accuracy rates, while nationality has no significant effect on accuracy rates, and the interaction between language level and nationality has no significant effect on frequency and diversity.

4　Research Results

This section examines the acquisition characteristics of prepositional frame structures among Thai, Vietnamese, and Indonesian learners at different proficiency levels from four dimensions: acquisition accuracy, error type distribution, structural usage frequency, and diversity. The study found that language level significantly affects acquisition accuracy, with elementary, intermediate, and advanced learners achieving accuracy rates of 54.79%, 76.70%, and 92.45% respectively, demonstrating a hierarchical developmental path of "form mastery—semantic distinction—pragmatic adaptation." Native language background produces differentiated effects on error types: Thai learners show prominent redundancy errors (23.80%), Vietnamese learners have the highest proportion of misordering errors (26.66%), and Indonesian learners display the most significant omission errors (52.94%), confirming the national characteristics of native language transfer effects. Furthermore, acquisition progress is mainly manifested in quality (accuracy) rather than quantity (usage frequency and diversity), with neither language level nor nationality having significant effects on the latter two.

5　Acquisition Strategy Recommendations for Prepositional Frame Structures

Based on the empirical research results of prepositional frame structure acquisition among Southeast Asian advanced Chinese learners, this section proposes optimization suggestions from two levels: acquisition strategies and teaching implications. For acquisition strategies, three effective pathways are summarized: first, mastering preposition usage through contextual understanding, where advanced learners accumulate perceptual knowledge through dialogue, reading, and writing examples, and can accurately select prepositions based on context; second, promoting acquisition through native language comparison, where Thai and Vietnamese learners utilize similarities in preposition-fronting typological languages, while Indonesian learners overcome interference from native language simplification features by strengthening metalinguistic awareness; third, promoting strategies targeting national differences, where Thai learners focus on structured input, Vietnamese learners emphasize semantic understanding, and Indonesian learners excel at systematic rule learning, with all three pointing to the importance of structure priority at the elementary stage, gradual gradient learning, and metacognitive monitoring. For teaching implications, three specific measures are proposed: designing differentiated content for different nationalities, such as "near-synonym preposition deletion games" for Thai students, "explicit framework" two-color marking method for Vietnamese students, and "locative word mandatory cloze method" for Indonesian students; stage-based teaching according to language proficiency, using "semantic diagram method" at the elementary stage, "native language comparison tasks" at the intermediate stage, and "discourse cloze tests" at the advanced stage; and achieving intelligent learning through smart technology, including developing Ollama model intelligent error correction plugins and building AntConc corpus learning modules, to provide personalized pathways and immediate feedback.

6　Conclusion

Based on the HSK Dynamic Composition Corpus, this study found that omission errors in prepositional frame structures almost disappear among Southeast Asian advanced Chinese learners, semantic errors decrease significantly, but redundancy errors remain stubborn; native language background significantly affects error types, with Thai learners showing prominent redundancy errors, Vietnamese learners displaying obvious misordering errors, and Indonesian learners mainly exhibiting omission errors. The study proposes three systematic learning strategies: contextual internalization, contrastive analysis, and differentiated pathways, and designs redundancy correction training, word order comparison exercises, and structural integrity enhancement programs for learners from the three countries respectively.