基于神经语义分析的档案智能开放鉴定模型构建探析  

Analysis of the Intelligent Archives Opening Appraisal Model Construction Based on Neural Semantic Analysis

作　　者：殷名 聂云霞[1] 李家霖 YIN Ming;NIE Yunxia;LI Jialin(Humanities School of Nanchang University,Nanchang,Jiangxi 330031)

机构地区：[1]南昌大学人文学院,江西南昌330031

出　　处：《档案学刊》2022年第2期46-55,共10页Academic Journal of Archives

基　　金：国家社会科学基金项目“新《档案法》背景下综合档案馆文化功能及其实现研究”(21BTQ089)研究成果。

摘　　要：神经语义分析同时具有神经网络模型的海量数据处理能力和自然语言处理技术的符号化特征,适用于解决档案开放待鉴定压力大、专业性强、精细化程度高等现实需求,能够开辟档案开放鉴定工作新图景,提高档案馆公共服务效能。从安全第一原则和多主体协同原则出发,构建基于神经语义分析的档案智能开放鉴定模型,以敏感词库赋能、以已鉴定档案训练、以多主体适时调整模型算法,能够实现对经过数据预处理的待鉴定档案的开放审核,建构全新的工作模式和完整的工作流程。Neural semantic analysis has both the ability of massive data processing of neural network model and the symbolic characteristics of natural language processing.It is suitable for solving the practical needs of high pressure,strong professionalism and high degree of refinement of the opening appraisal of archives,which can open up a new prospect for the appraisal of archives opening and improve the effectiveness of public services of archives.Starting from the principle of safety first and multi-agent cooperation,an intelligent archives opening appraisal model based on neural semantic analysis is constructed.With the empowerment of sensitive thesaurus,the training of appraised archives and the timely adjustment of model algorithm by multi-agent,the audit of opening archives to be appraised after data preprocessing is performed,and a new working mode and workflow are constructed.

关 键 词：档案开放鉴定 神经语义分析 智能 算法分析 模型构建 

分 类 号：G274[文化科学—档案学]