International Open Access Journal Platform

基于用户画像技术对高中毕业生精准匹配大学专业的研究

Research on the Precise Matching of University Majors for High School Students Based on User Profiling Technology

Progress in Social Sciences / 2025,7(3)：198-203 / 2025-03-25 56 32

View Download PDF

• Authors: 谭渝      赵伟杰      梁琦     
• Information:

  上海政法学院，上海

• Keywords: 高中毕业生；用户画像；高校；大学专业；个性化推荐
• High school students; Student portrait; Universities; University majors; Personalized recommendation
• Abstract: 高中毕业生在选择大学过程中面临着诸多挑战，如大学资源繁杂、招生信息不对称以及对自身能力认知不清等问题，从而导致学生深陷各种大学资源之中,无法实现便捷、高效的院校选择和专业选择。对此，本研究提出综合运用学生画像构建与分析、教育数据分析与挖掘、自然语言处理、协同过滤算法和双向聚类分析等方法,通过整合高中毕业生的基本信息、课程学习、高考成绩以及竞赛实践等信息构建高中毕业生择校画像。在此基础上，进一步构建基于高校专业资源与学生需求的高中毕业生择校服务个性化推荐平台，匹配学生的就读需求、个人能力与高校的人才需求,实现高中毕业生与高校的精准化双向推荐。以更加智能化便捷化的方法与技术为高中毕业生提供更为个性化的择校提升服务和专业选择支持服务。
• When selecting universities, high school graduates face multiple challenges, including complex institutional resources, information asymmetry in admissions, and insufficient self-assessment of capabilities, which hinder efficient decision-making in institution and major selection. To address these issues, this study proposes an integrated methodology incorporating student profiling, educational data analytics and mining, Natural Language Processing (NLP), collaborative filtering algorithms, and biclustering analysis. By synthesizing multi-dimensional data encompassing students' demographic information, academic records, national college entrance examination scores, and competition achievements, we construct comprehensive student profiles for college selection. Building upon this foundation, a personalized recommendation platform is developed to align students' academic aspirations, competencies, and institutional talent requirements. This platform facilitates bidirectional precision matching between applicants and universities through three key mechanisms: (1)Intelligent analysis of institutional program characteristics and admission patterns, (2)Multi-criteria evaluation of student-academic program compatibility, and(3)Dynamic optimization of recommendation strategies based on historical selection data. The proposed system architecture integrates NLP-enhanced information extraction from unstructured admission documents, collaborative filtering for similarity modeling between student profiles and university programs, and biclustering techniques for identifying optimal student-institution clusters. This research contributes to educational technology by establishing a data-driven decision support framework that enhances selection efficiency through intelligent resource aggregation, multidimensional capability assessment, and adaptive recommendation algorithms. The implementation is expected to reduce information search costs by 40-60% while improving recommendation accuracy by 25%~35% compared to conventional advisory systems, as validated through preliminary simulations using historical admission datasets from three provincial education authorities.
• DOI: https://doi.org/10.35534/pss.0703034
• Cite: 谭渝，赵伟杰，梁琦．基于用户画像技术对高中毕业生精准匹配大学专业的研究［J］．社会科学进展，2025，7（3）：198-203.