基于深度学习的网络档案开发与管理方法  

The Development and Management Method of Network Archives Based on Deep Learning

作　　者：赵容 ZHAO Rong(Chongqing Vocational Institute of Tourism,Chongqing 409000,China)

机构地区：[1]重庆旅游职业学院,重庆409000

出　　处：《信息与电脑》2021年第11期36-38,共3页Information & Computer

摘　　要：为了降低由分类不准确引起的档案管理效率低问题,笔者提出一种基于深度学习的网络档案开发与管理方法。首先,对基于Bootstrapping技术的语料库方法进行改进,在迭代过程中引入信息元评价机制提升挖掘质量;其次,利用长短期记忆模型在语料库抽取多类型细粒度的档案信息元;最后,建立AlexNet网络,对档案进行分类管理。实验结果表明,本文所提方法对目标档案的查询结果准确率为100%,并且对单独档案的查询时间基本在5s以内,具有良好的管理效果。In order to reduce the low efficiency of archives management caused by inaccurate classification,this paper proposes the research of network archives development and management method based on deep learning.Firstly,the corpus method based on bootstrapping technology is improved,and the evaluation mechanism of pattern and information element is introduced in the iterative process to improve the mining pattern and mining quality;Secondly,the long-term and short-term memory model is used to extract multiple types of fine-grained archival information elements from the corpus;Finally,the alexnet network is established to manage the archives by classification.The experimental results show that the accuracy of the proposed method is 100%,and the query time for individual files is basically less than 5 s,which has good management effect.

关 键 词：深度学习 网络档案 档案信息元 AlexNet网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]