为提高档案管理的效率和准确性,提出了一种基于智能优化深度网络的档案数据分析方法.该方法结合了Transformer网络的特征提取能力和灰狼优化算法(Grey Wolf Optimizer,GWO)的参数优化能力,显著提高了档案管理的效率和准确性.将提出的方法运用到档案数据的多分类任务中,实验结果表明,通过GWO算法优化后的Transformer-GWO模型在与xgboost分类模型结合时取得了最佳性能,其宏观精确度、宏观召回率和宏观F1分数分别达到0.893、0.878以及0.885,提出的方法有效提升了档案管理的智能化水平.
The archive data analysis method based on intelligent optimization of deep networks
To improve the efficiency and accuracy of file management,this paper proposes an archive data analysis method based on intelligent optimization deep network.This method combines the feature extraction ability of Transformer network and the parameter optimization ability of grey Wolf optimization algorithm(Grey Wolf Optimizer,GWO),which significantly improves the efficiency and accuracy of file management.The proposed method was applied to the multiple classification task of archival data,and the experimental results showed that the Transformer-GWO model achieved the best performance when combined with the xgboost classification model,and the macro accuracy,macro recall rate and macro F1 score reached 0.893,0.878 and 0.885 respectively.The proposed method effectively improves the intelligent level of archives management.
archive dataTransformerGrey Wolf Optimization algorithmclassification
万方数据
维普
