摘要
提出一种基于深度学习的PDM文档自动审核算法,该算法根据专家产生的文档训练集进行学习训练以形成模型参数,使用产生的模型参数对新文档进行审核以发现其中存在的各类错误,为审核人员提供辅助支撑。对用到的PDM文档分词、语句表示及分类进行了描述。该算法在客观性、经济实用性和即时性等方面都优于传统人工审核,为海量PDM文档的审核管理提供了一种解决途径,为以后PDM文档的智能化处理奠定了基础。
A PDM document automatic examination algorithm based on deep learning is proposed. It uses the document training set produced by experts for learning and training to generate model parameters, and uses these model parameters for new document examination to discover the various errors and provides assistant support for the examination personnel. The PDM document word segmentation, sentence expression and classification are described. The method is better than conventional manual examination in objectivity, economical practicability and timeliness. It provides a solution for the examination management of massive PDM documents, and provides a reference for the intelligent processing of PDM documents.
作者
宁凌
NI Ling(The 54th Research Institute of CETC,Shijiazhuang Hebei 050081,China)
出处
《计算机与网络》
2018年第10期57-58,61,共3页
Computer & Network
