基于深度学习的档案辅助分类应用研究
Research on the Application of Archive-assisted Classification Based on Deep Learning
房开乾 1王鑫鑫 1欧叶玉1
作者信息
- 1. 中国工程物理研究院档案馆,绵阳 621900
- 折叠
摘要
本文通过搭建直接训练分类模型和预训练分类模型,对文本特征进行深层次抽象和学习,实现文本材料的自动识别与分类.实验选取不同规模数据集作为训练和测试样本,将调整优化并在公开数据集上训练验证过的直接训练模型和预训练模型应用于真实档案文本分类,取得了较高的准确率.
Abstract
In this paper,a direct training classification model and a pre-trained classification model are built to abstract and learn text features at a deep level,so as to realize the automatic recognition and classification of text materials.In the experiment,datasets of different sizes were selected as training and test samples,and the direct training model and pre-trained model that were adjusted and optimized and trained on the public dataset were applied to the text classification of real archives,and a high accuracy was obtained.
关键词
人工智能/深度学习/档案分类Key words
Artificial Intelligence/Deep Learning/Classification of Archives引用本文复制引用
基金项目
国家档案局科技项目(2023-X-034)
出版年
2024
NETL
NSTL
万方数据