人工智能图书采访系统的学习训练集构建与维护  

Construction and Maintenance of Learning Training Set for Artificial Intelligence in Book Interview

作　　者：原小玲[1] 王红[2] YUAN Xiaoling;WANG Hong(Taiyuan Science and Technology University,Taiyuan 030024;Shanxi University of Finance and Ecnomics,Taiyuan 030006,China)

机构地区：[1]太原科技大学图书馆,山西太原030024 [2]山西财经大学图书馆,山西太原030006

出　　处：《晋图学刊》2020年第3期22-25,共4页Shanxi Library Journal

基　　金：国家社会科学基金项目“人工智能图书采访决策模型研究”(项目编号:17BTQ026)。

摘　　要：学习训练集是人工智能采访系统进行深度学习训练的依据。学习训练集数据主要包括馆藏数据、用户数据及其他相关数据。这些数据以主题词数据为核心,互相建立关联,作为人工智能采访系统的学习训练素材。学习训练集数据运行要在优先级与阈值设置基础上进行。学习训练集数据维护要着重于数据的长期性与持续更新,并注意其不完整性带来的局限。The learning training set is the basis for the in-depth study of artificial intelligence in book interviews.Learning training set data includes historical data and other relevant data.These data establish relationships with each other as a material for neural network learning training based on the subject word data.Learning training set data run on the basis of priority and threshold setting.Data maintenance should focus on the long-term and continuous updating of data,and pay attention to the limitations caused by its incompleteness.

关 键 词：学习训练集 馆藏数据 用户数据 其他数据 数据关联 

分 类 号：G253.1[文化科学—图书馆学]