一种基于爬山算法的学科领域热点探索模型

A Hotspot Exploration Model in Discipline Field Based on Mountain-climbing Algorithm

[目的/意义]热点发掘一直是图书馆领域的重点研究内容之一,通过共引和共词分析探测学科领域热点都存在一定的局限性,针对存在的问题,借鉴爬山法进行改进研究。为适应大数据时代的热点挖掘问题,将计算机算法与文献关键词分析相结合,构建研究热点分析模型,探索分析研究热点的方法,对于学科领域研究热点具有重要意义。[方法/过程]采用改进的Apriori-Hill-Mount爬山算法,拓展关键词共现分析和聚类方法,构建了篇与篇之间的关键词的关联规则挖掘模型;通过实证研究科学计量学及图书情报学两个领域,论证该模型的可行性。[结果/结论]通过对上述两个领域的文献进行篇-篇关键词强关联规则挖掘和分析,分别得到各自领域的的研究热点;其结果与采用Citespace及Spss呈现的热点有很高的吻合度。但该算法揭示的信息更全面、更直观地呈现研究主题热点,对学科热点探索体系方法及手段具有一定的补充及辅助作用。该方法是在一个局部范围内进行搜索,且阈值的设定也具有人为因素,因此得到的可能是局部的优化,其更大范围的可适性有待进一步论证。

[Purpose/Significance]Hotspot mining has always been one of the key research contents in the field of library. There are some limitations in detecting hotspots through co-citation and Co-word analysis. In view of the existing problems, the mountain climbing method is used in order to improve the research analysis. To solve the hotspot mining problems in the era of big data, it is of great significance to combine computer algorithms with keyword analysis in literature, to construct a hotspot analysis model and to explore new methods of hotspot analysis.[Method/Process]By adopting the improved Apriori-Hill-Mount climbing algorithm, expanding keyword co-occurrence analysis and clustering method, an association rule mining model of the keywords between the texts was constructed. The feasibility of the model was verified through the empirical study of the two subjects of Bibliometrics and Library and Information Science. [Result/Conclusion]Through the mining and analysis of the strong association rules of keywords in the above two fields, the research hotspots were obtained respectively. The results have a high coincidence and similarity with the hot spots or clusters presented by Citespace and Spss, but the information revealed by this model is more comprehensive and more intuitive. It is supplementary and helpful to the research methods and means of discipline hotspots.The method is to search within a local range, and the threshold setting also contains human factors, so the result obtained may be a local optimization. The greater scope of its suitability needs to be further explored.