摘要
[目的/意义]人才视角下明确高科技企业的专业技能需求以及变化趋势,有助于加快聚合创新要素、逐步组建技术团体以及建立健全协作体系。[方法/过程]文章以IC(Integrated Circuit)领域在职者简历信息为数据基础,基于时间维度下的协同过滤推荐算法,实现领域企业和专业技能间的关系预测和评估;结合Word2vec词向量模型与LDA模型,对简历信息进行数据挖掘与语料库扩充,进而识别技能主题;随后刻画技能主题演化路径并进行可视化表达,明确IC领域技能主题发展现状及演化规律。[结果/结论]研究表明,基于时间维度下的协同过滤思路适用于企业与技能关系预测,平均准确率达96.68%;IC领域的技能主题演化呈现“聚合→分散→再聚合”的趋势,各技能主题路径演化规律有所差异,具体表现形式为螺旋交叉演化、创新迭代演化和技术依附演化;研究给出的IC行业整体技能发展态势和路径变化规律,可为高科技企业和在职者提供更精准的检索方法和技能选择方向。
[Purpose/significance]To clarify the professional skill demand and change trend of high-tech enterprises from the perspective of supply and demand optimization is helpful to accelerate the aggregation of innovation elements,gradually establish technical groups and establish a sound cooperation system.[Method/process]Based on the resume information of incumbents in the Integrated Circuit(IC)field,the collaborative filtering recommendation algorithm in the ime dimension was used to predict and evaluate the relationship between enterprises and professional skills in the field.By combining Word2vec word vector model and LDA model,data mining and corpus expansion of resume information are carried out to identify skill topics.Then describe the evolution path of skill theme and carry out visual expression to clarify the development status and evolution law of skill theme in IC field.[Result/conclusion]The research shows that the collaborative filtering idea based on time dimension is suitable for the prediction of the relationship between enterprises and skills,with an average accuracy of 96.68%.The evolution of skill theme in IC field shows the trend of"aggregation→dispersion→reaggregation",and the evolution law of each skill theme path is different,which is embodied in spiral cross evolution,innovation iterative evolution and technology dependency evolution.The overall skill development trend and path change law of IC industry given by the study can provide more accurate retrieval methods and skill selection directions for high-tech enterprises and incumbents.
作者
窦路遥
黄运聪
白增亮
李毅
周志刚
DOU Luyao;HUANG Yuncong;BAI Zengliang;LI Yi;ZHOU Zhigang(National Science Library(Wuhan),Chinese Academy of Sciences,Wuhan 430071;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100080;School of Information Technology,Shanxi University of Finance and Economics,Taiyuan 030006)
出处
《科技情报研究》
CSSCI
2024年第4期114-126,共13页
Scientific Information Research
基金
国家自然科学基金项目“多源数据融合场景下的对抗式隐私洞察靶向保护技术研究”(编号:61902226)。
关键词
供需优化
时间维度
协同过滤
关系预测
路径演化
IC领域
supply and demand optimization
time dimension
collaborative filtering
relationship prediction
path evolution
IC field
