联系我们

论文查询结果	返回搜索

论文编号：	16162
作者编号：	2320234200
上传时间：	2026/6/10 14:46:21
中文题目：	基于学科会聚的保研专业信息智能推荐系统构建研究
英文题目：	Research on the Construction of an Intelligent Recommendation System for Postgraduate Admission Major Information Based on Disciplinary Convergence
指导老师：	冯湘君
中文关键字：	学科会聚；人工智能；知识计量学；智能推荐系统：
英文关键字：	Disciplinary convergence;Artificial Intelligence; Knowledge Metometrics; Intelligent recommendation systems;
中文摘要：	“四新”建设是学科交叉的里程碑事件，人工智能作为“四新”建设核心学科，交叉会聚特征日益显著。对学科会聚演化分析、学科会聚规律总结及实践运用进行系统性探索具有重大意义。本研究以人工智能学科为例，以中国知网2018-2025年间包含“TP18”的多分类号文献共51470条为研究对象，综合运用文献计量学、网络科学理论等，对人工智能学科会聚现象进行剖析，总结会聚规律并预测未来发展。此外，以关乎学科会聚的保研专业信息为研究对象进行智能推荐系统构建和开发，并完成原型设计。完成学科会聚的“描述－解释－预测－应用”的研究闭环，为其他学科会聚研究提供参考。 在研究方法上，本研究构建了“静态结构分析-动态时序分析-量化预测建模-系统实现验证”四阶体系。首先，运用Python的Pandas等科学计算库，对原始文献数据进行深度清洗，构建了以中图分类号为节点、共现关系为边的加权无向网络，利用VOSviewer进行了静态网络分析，识别出与TP18相关的学科聚类。其次，借助CiteSpace软件进行动态文献计量分析，并利用内置突现检测算法识别了不同时间段的研究热点。接着，系统计算了度中心性等指标，刻画了各学科节点在网络中的影响力。研究引入Granovetter的弱连接理论，识别出高创新潜力的学科交叉方向，对人工智能学科发展趋势进行预测。 最后，将前述分析的理论成果，转化为具体的算法模块与功能设计，以Java、Python等编程语言，设计并实现了面向保研场景的专业信息智能推荐系统（或简称信息系统）原型，完成了从理论发现到实践应用的完整闭环。 本研究构建并验证了一个学科会聚系统分析框架，并用实践成果验证理论分析，形成研究闭环。学科会聚的理论研究有助于高校优化学科布局，实践上有助于保研学子发现适合的交叉专业。研究同时指出，未来可在纳入多元数据、研发面向不同群体的智能推荐系统等方面进行完善。本研究深化了对人工智能学科交叉规律的理解，在一定程度上推动学科交叉研究成果落地转化。图41个，表15个，参考文献54篇。
英文摘要：	The "Four New" initiative represents a milestone in interdisciplinary integration, with artificial intelligence—as the core discipline within it—exhibiting increasingly pronounced convergent characteristics. Systematic exploration of disciplinary convergence evolution analysis, pattern identification, and practical applications holds significant importance. This study takes the field of artificial intelligence as an example, analyzing 51,470 multi-classified literature entries from China National Knowledge Infrastructure (CNKI) covering the period 2018–2025 that contain the tag "TP18." By integrating bibliometric methods and network science theories, the research examines AI disciplinary convergence phenomena, summarizes underlying patterns, and forecasts future trends. Additionally, the study develops an intelligent recommendation system based on graduate admission program information related to disciplinary convergence, completing a closed-loop research cycle encompassing description, explanation, prediction, and application, thereby providing reference for similar convergence studies across disciplines. In terms of research methodology, this study establishes a four-tier framework comprising "static structural analysis—dynamic temporal analysis—quantitative prediction modeling—system implementation validation." First, using scientific computing libraries such as Python's Pandas, the raw literature data were subjected to thorough cleansing, resulting in a weighted undirected network where Chinese Library Classification codes served as nodes and co-occurrence relationships as edges. Static network analysis was conducted with VOSviewer to identify disciplinary clusters associated with TP18. Subsequently, dynamic bibliometric analysis was performed using CiteSpace software, employing its built-in emergence detection algorithm to pinpoint research hotspots across different time periods. The study then systematically calculated metrics such as degree centrality to characterize the influence of each disciplinary node within the network. By applying Granovetter's weak link theory, it identified interdisciplinary directions with high innovation potential and predicted future trends in artificial intelligence research. Finally, the theoretical findings from the aforementioned analysis were translated into concrete algorithmic modules and functional designs. Using programming languages such as Java and Python, a prototype of an intelligent professional information recommendation system (or simply an information system) tailored for graduate admission scenarios was developed and implemented, thereby completing the full closed-loop process from theoretical discovery to practical application. This study established and validated a systematic analytical framework for disciplinary convergence, using practical outcomes to validate theoretical analyses and thereby forming a closed research loop. The theoretical research on disciplinary convergence aids universities in optimizing their disciplinary structures, while its practical implications assist students eligible for direct admission to graduate programs in identifying suitable interdisciplinary fields. The study also highlights potential future improvements, such as incorporating diverse data sources and developing intelligent recommendation systems tailored to different user groups. This work deepens our understanding of the patterns governing interdisciplinary integration in artificial intelligence and facilitates the practical application of interdisciplinary research findings. The study includes 41 figures, 15 tables, and references to 54 publications.
查看全文：	预览  下载(下载需要进行登录)