Subjects: Library Science，Information Science >> Library Science submitted time 2023-10-08 Cooperative journals: 《知识管理论坛》
Abstract： [Purpose/significance] The analysis of technological topic evolution is of great significance for understanding the technological development process and forecasting technology trends in the future. Exploring the roadmap of research methods of technological topic evolution analysis helps is helpful to understand its research status and lay the foundation for further research. [Method/process] This paper divided the existing research methods of technological topic evolution analysis into three major categories: qualitative methods, quantitative methods and hybrid methods combining qualitative and quantitative analysis, and then describes representative methods for each category. [Results/conclusion] Based on the systematic analysis of current research methods of technological topic evolution, the author discusses the shortage of existing research methods and prospects for future research.
Subjects: Library Science，Information Science >> Library Science submitted time 2023-07-26 Cooperative journals: 《图书情报工作》
Abstract： [Purpose/significance] To study the influence of time factor on patent cited frequency can reduce the restriction of time factor on technical evaluation activities and improve the accuracy and reliability of evaluation.[Method/process] This paper collected U.S. patent data from 1975 to 2017, and carried out the revision study of patent cited frequency based on fixed effect method. The patents were grouped according to different publication years and different technical fields. Group mean and six TOP quantiles were selected as the benchmarks of patent cited frequency, and the benchmarks of patent cited frequency for both the current time point and the historical time series were counted. Then a neural network model was established to fit the timing variation of the benchmarks, thus predicting the benchmarks of future statistical time points.[Result/conclusion] The time difference between publication years and statistical years of patents makes it impossible to directly compare patent citations. This paper establishes benchmarks for patent citations based on different technical fields, different publication years and different statistical years, providing reference for patent evaluation.
Subjects: Library Science，Information Science >> Information Science submitted time 2017-11-08 Cooperative journals: 《数据分析与知识发现》
Abstract：【目的】利用Word2Vec 深度学习技术从面向大众的健康信息中寻找疾病关联, 解决非医学人士通常不了 解多种疾病之间存在的关联, 从而影响到健康信息搜寻中的全面性和有效性的问题。【方法】由专家选取30个常见疾病主题,从高质量医学新闻网站上采集对应疾病的文档, 运用Word2Vec技术对各疾病的相关文档构造词 向量,计算向量距离判断疾病关联。通过与专家评分的相关分析衡量判断结果的准确性。【结果】优情况下,Word2Vec 得到的结果与专家评分相关系数达到0.635。通过对比不同的算法模型、优化方法、数据规模及重要 参数对结果的影响,发现Skip-Gram模型结合负样本数为20的Negative Sampling优化方法在大规模数据集上的 实验结果优。【局限】疾病主题选取宽泛时, 影响Word2Vec 判断准确性, 本文的疾病主题选取粒度有待改善。【结论】利用 Word2Vec 技术在面向大众的健康信息源中也可以探测疾病关联, 其有效性表明该技术可用于改善 大众的健康信息搜寻的个性化服务。
Hits
YES
Hits
Downloads
Comment