数据业务语义自动识别模型的构建与应用  

Construction and application of automatic recognition model for data business semantics

作　　者：苏志勇 朱艺媛 陈伟 曾荣甫 何秋芸 SU Zhiyong;ZHU Yiyuan;CHEN Wei;ZENG Rongfu;HE Qiuyun(State Network Xintong Yili Technology Co.,Ltd.,Xiamen 361001,Fujian China;Fujian Rhodes Digital Technology Co.,Ltd.,Fuzhou 350200,China)

机构地区：[1]国网信通亿力科技有限责任公司,福建厦门361001 [2]福建罗德数字科技有限公司,福建福州350200

出　　处：《粘接》2025年第1期145-148,共4页Adhesion

摘　　要：针对数据业务语句的词语解析和信息抽取存在较大差异的问题,研究一种基于自然语言处理(NLP)技术的语义自动识别系统。在本识别系统的功能基础上,加入了基于卷积神经网络(CNN)算法的人工智能模块,对所接收的信息数据语义进行卷积和池化,使数据业务语句解析和信息抽取的差异问题有很大改善。为了提高语义识别效率,进一步采用分词法(WS),来对接收的语句进行分词解读,达到降低词语歧义的目的,减少对整体语义识别的影响。试验结果表明,通过本系统对数据业务的语义识别精准度在90%以上,表明该系统对解决语义识别问题具有较强的实用性和优越性。Aiming at the big difference between word parsing and information extraction of data business state⁃ments,an automatic semantic recognition system based on Natural Language Processing(NLP)technology was studied.On the basis of the functions of the recognition system,an artificial intelligence module based on Convolu⁃tional Neural Networks(CNN)algorithm was added to convolve and pool the semantics of the received information data,which greatly improves the difference between data business statement parsing and information extraction.In order to improve the efficiency of semantic recognition,Word segmentation(WS)was further adopted to interpret the received sentences,so as to reduce word ambiguity and reduce the impact on the overall semantic recognition.The test results showed that the accuracy of semantic recognition of data business through this system was more than 90%,which indicates that this research system has strong practicability and superiority in solving semantic rec⁃ognition problems.

关 键 词：语义识别 NLP技术 CNN算法 分词法 人工智能 

分 类 号：TP317.2[自动化与计算机技术—计算机软件与理论]